the cult of sulis-minerva at bath - researcharchive.vuw.ac.nz
TRANSCRIPT
The Cult of Sulis-Minerva at Bath: The Religious Ritual of the Patron Goddess at
Bath.
By
James McBurney A thesis submitted to the Victoria University of Wellington in fulfilment of the
requirements for the degree of Master of Arts in Classical Studies 2016
1
Table of contents:
Acknowledgements…………………………………………………………………………..2
Abstract………………………………………………………………………………………3
Intoduction…………………………………………………………………………………...5
1.Bath and Sulis’ temple complex………………………………………….........................11
The history of Aquae Sulis…………………………………………………......................12
Dating Bath’s defixiones………………………………………………………………….21
The use of metal at Bath………………………………………………………………….28
When Roman Bath ended………………………………………………………………...33
Conclusion…………………………………………………………………………………..44
2.Prayers or Curses? A Study of Bath’s Defixiones………………………………………45
Vulgar Latin on Bath’s defixiones………………………………………........................47
Defixiones from around the Ancient World……………………………........................52
The power of prayer in the Roman World……………………………………………..56
The formulae and subject of defixiones…………………………………………………61
Differences between Bath’s defixiones and examples from the Mediterranean………………………………………………………………………64
The relationship between defigentes and deities on curse tablets………...............................................................................................................75
The defigentes of Bath……………………………………………………........................83
Conclusion………………………………………………………………………………...85
3. The Cult of Sulis-Minerva and the goddess’ divine attributes………………………...87
The lack of evidence regarding cult activity at Bath…………………………………...87
The Romanization of ancient British religion…………………………………………..91
The structure and layout of Bath’s Romano-Celtic temple complex…........................95
Bath’s Gorgon pediment: A hybridization of Roman and Celtic religion……………………………………………………………………………….99
Sulis’ Minervan attributes: The conflation of two goddesses…………........................103 A possible connection to the sun: Sulis’ solar attributes……………………………....108 Sulis as a judicial deity…………………………………………………………………...111 Religious evidence from Bath’s sacred spring………………………….........................113 Conclusion………………………………………………………………………………...119 Conclusion………………………………………………………………….........................121 Bibliography……………………………………………………………………………….126
Figure List………………………………………………………………………………….137
2
Acknowledgements.
To my supervisor Judy Deuling, who met with me on a weekly basis and constantly kept me on my feet. Thank you for helping me to develop my thesis over the past two and a half years
as well as providing enthusiastic advice. I’m going to miss meeting with you at 8:30 am Friday mornings. I appreciate everything you have done for me.
To my other supervison Diana Burton who has provided invaluable direction and advice. Your ideas helped keep my thesis on track and stopped me from wondering off course. Your
insight has helped me to develop my ideas and get them down on paper. Thank you for working with over the past two years.
I would like to thank my family: Dad, Jess and Max for feeding me and making sure that I’m alive. I would also like to thank my cat Iggy for being a living stress ball.
I would also like to thank the OK518 team: Nikki Carter, Julia Simons, Richard Buckley-Gorman, Laura Corn, Tim Smith and Cameron Stanton. Thank you for supporting me and
giving me advice when times were at their hardest. Praise Solon!
Lastly I would like to thank The Lads: Luke van den Hoven, Sylvan Haeusler and Tobin Haeusler. You have been there from the beginning and provided laughs and pints when I
needed them the most.
3
Abstract. The Roman City of Bath, also known as Aqua Sulis, lies in the modern British county of Somerset in the south-east of England. During the Roman occupation of ancient Britain, Bath became a significant Roman town centred on a large religious complex. As the Roman city lies underneath the modern city Bath, excavation of both the temple complex has been difficult. To add further problems, Bath was only mentioned in one ancient source, Solinus. Consequently, there is a large gap in the knowledge we have about Roman Bath and its patron goddess. As such a large Romano-Celtic temple complex, Sulis’ cult has important contributions to religion in Roman Britain. Subsequently, studying and understanding Sulis’ cult is important to the study of Roman Britain. This thesis discusses features of Sulis’ cult and what this may tell us about the goddess’s attributes as well as how her cult functioned.
The large Romano-Celtic temple was functional from c.65 to c.400 CE. However, there is evidence which would suggest that Sulis was worshipped by the ancient Britons before the Romans had a permanent presence in Britain. This thesis will place Roman Bath within the wider context of Romano-British history, outlining how it functioned through architecture and evidence for the temple’s gradual decline.
Scholarship has agreed that Sulis is a Celtic deity who was worshipped by the Celts before the Roman arrived in Britain. Through Roman religious sensibilities, Sulis was conflated with the Roman goddess Minerva. Most of the physical remains at Bath are architectural features, votive offerings and altars. Many links have been drawn between Sulis and her thermal spring. For example, Sulis-Minerva has been regarded as an important healing divinity and her temple complex a place people can go for healing. This thesis will discuss Sulis and her connection to the goddess Minerva as well as what the goddess’s relationship was to the Romans and Britons.
Attention has been drawn to a large cache of 130 Latin defixiones, or curse tablets, discovered in Sulis spring. The curses most commonly beseech Sulis to hunt down a culprit and punish them. The defixiones constitute an important source of evidence regarding Sulis’ attributes. There has been some debate as to the nature of these curse tablets as there have been suggestions that they read more as ‘prayers for justice’. This thesis will explore the idea that the tablets acted as a medium for a devotee to ask the goddess for retribution against a perceived wrong. A comparison will be drawn between Bath’s curse tablets and other forms of Roman prayers comparing the two. As of now, the defixiones contribute a large portion of evidence towards religious life at Bath.
Sulis represents hybridization between two ancient civilizations. On the one hand, Sulis had strong roots to ancient British religion but after Roman occupation her cult became predominantly Roman in form. I will discuss the remaining aspects of Celtic religion at Bath, such as in the Gorgon pediment, and how this was changed under Roman rule.
5
Introduction
The city of Bath, or Aquae Sulis in Latin, is one that has been occupied constantly throughout
its existence. During the Roman occupation of ancient Britain, Bath became an important
Roman town centred on a large religious complex. Modern Bath lies in the southwest of
Britain in the county of Somerset. Today, a large modern town is built where the ancient city
once sat. As one of Britain’s only thermal hotspots, Bath has continued to draw people to its
modern spa complex and still continues to be used today. At the heart of Bath was the temple
of the goddess Sulis Minerva. As a goddess, Sulis is mysterious. Bath and its patron goddess
do not appear in ancient literature, making the study of the goddess and her temple complex
difficult. The only evidence available to scholars has been the physical evidence left behind
at the Roman temple. However, the presence of the modern city of Bath has made excavation
of the ancient city problematic. Even so, the size of Bath and the intricacy of Sulis’ large
temple complex highlights a place that was very important to the Britons and the Romans
living in Bath. The large Romano-Celtic temple of Sulis Minerva was elaborately decorated
and functioned from c.65 CE- c.400 CE. The temple’s aggrandizement points to a religious
complex that was popular as well as wealthy. Consequently, it is peculiar that we know so
little about Bath and the goddess that resided there. As such a prominent Romano-British
city, Bath is important in the study of Roman Britain. Most physical remains at Bath are
architectural features, votive offerings and altars. Many links have been drawn between Sulis
and her hot spring. As hot springs were important places of healing in the ancient world, it
has been suggested that Sulis acted as a healer through the properties of her spring.1
1 Green (1995) 93.
6
Bath and Sulis’ cult has been the focus of large amounts of scholarship. However,
conclusions have been limited by a lack of definitive evidence. There are still many gaps in
scholarship about the goddess Sulis Minerva. A lack of evidence has led to speculation about
the goddess Sulis and her cult. Although Barry Cunliffe has compiled a comprehensive
archaeological report on Bath in his two volumes of The Temple of Minerva Sulis at Bath,2
most scholarship such as Miranda Green has tended to focus on only a few aspects of Sulis’
cult such as the Gorgon pediment or Bath’s hot springs. Most scholarship has agreed that
Sulis is a prominently Celtic deity who was worshipped by the Celts before the Romans
occupied Britain.3 Through her connection to the thermal hot springs, scholars such as
Miranda Green have emphasized an important healing aspect of Sulis’ cult at Bath.4
Furthermore, Cunliffe has the most detailed discussion of Bath’s defixiones in his publication
on the region’s archaeology.5 Cunliffe’s work focuses on all aspects of the defixiones with
translations and notations considering the inscribed curses. Aside from Cunliffe’s
consideration, most scholarship regarding Bath is made within a larger discussion of Bath and
the goddess Sulis. However, Henk Versnel uses Bath’s defixiones as examples concerning
‘prayers of justice’ in the ancient world. Versnel provides an alternative consideration as to
the working of defixiones in the ancient world. As of yet, there is very little scholarship that
specifically discusses Sulis and her defixiones. The defixiones constitute an important source
of information regarding Sulis’ divine attributes. The defixiones are small inscribed pieces of
thin pewter and lead. The curses most commonly beseech Sulis to hunt down a specific
culprit and punish them. Interestingly, the curses are commonly made in response to the theft
2 Cunliffe and Davenport (1985), Cunliffe (1988). 3 Cunliffe (1988), Green (1996), Mattingly (2006), Henig (2004). 4 Green (1991) 20. 5 Cunliffe (1988) 59-269.
7
of an item. Although Bath curse tablets share similarities with curse tablets from around the
ancient world, they also differ in a number of ways. In most cases a curse tablet is employed
to move the odds of a particular event such as sport into one’s favour (at the expense of
another person) through supernatural means.
Bath’s defixiones have led to much speculation about Sulis’ attributes as a goddess. The sheer
number of tablets found in the spring indicates that the use of defixiones was an important
part of Sulis’ cult. Importantly, Bath is currently one of only two sites which have yielded
such a large number of curse tablets, with the other being Uley. Because of the abnormalities
of Bath’s defixiones, alternative considerations have been given to the curse tablets. Scholars
such as Versnel have suggested that the curses are actually a form of ‘prayer for justice.’ A
devotee can approach Sulis and ask her for restitution for theft. This contrasts greatly with the
curse tablets from the greater ancient world. Defixiones such as the ones found at Bath are
made in retaliation to a perceived wrong; the defixiones’ dedicators are looking for restitution
through the goddess. The tablets are important pieces of evidence when studying an
ambiguous goddess such as Sulis. Furthermore, these forms of prayers are unique to ancient
Britain. As of yet, there are no defixiones from England that match their Mediterranean
cousins. This thesis will seek to explain the goddess Sulis-Minerva’s role as deity within
Romano-British religion and her importance within Bath’s sacred temple complex. I will
discuss Sulis’ divine attributes through the predominantly material remains and emphasize
her role as a judicial deity, capable of crippling her devotee’s enemies while also being a
powerful healing force due to her sacred spring. This thesis will also examine sulis’ identity
as both a Roman and British deity.
8
In my first chapter, I will discuss the site of Bath and place it within a historical context.
Currently, we are not able to put a certain date on Bath and when it reached its most
prominent period as a Roman temple complex. I will explain how the city Bath functioned
and when it was at its most active. I will discuss Bath’s role in Roman Britain and what kind
of town Bath was. I will examine what Bath has in common with other Roman towns as well
as what makes it unique. Furthermore, by examining other parts of Bath we can place a
rough date on the deposition of the temple’s defixiones. The palaeography of the defixiones as
well as coin hoards place the deposition of Bath’s defixiones in the third and fourth century
CE. I will also discuss when and how Bath ceased to function as a Roman temple complex.
Several archaeological finds help to outline Bath’s derelictions temple. It is likely that Bath
was affected by the same collapse as other parts of the ancient world. Consequently, Bath
had a short life span from the height of its power to when it ended.
My second chapter is a study of Bath’s defixiones. I will outline what constitutes a typical
curse tablet in the ancient world and how they evolved and functioned. I will compare Bath’s
own defixiones to those from the ancient Mediterranean and talk about their similarities and
differences. Furthermore, I will explain the place of defixiones in ancient Britain and how
they match up with Bath’s own curse tablets. Through the variances in Bath’s defixiones I
will discuss the possibility of them being ‘prayers for justice’, as Versnel contended. This
chapter will also discuss Roman prayer, how it is used and how this is similar to the language
on Bath’s defixiones. Through discussing prayer, I will highlight the thin line between
‘magic’ and religion in the ancient world. This chapter will highlight the judicial nature of the
defixiones and how this reflects the relationship between Sulis and her devotees. I will also
discuss the defixiones as a medium to communicate with Sulis in her spring. Through the
defixiones, Sulis likely acted as a form of supernatural law enforcement. Furthermore, this
9
chapter will also discuss the dedicators, or defigentes, of the curse tablets. It is likely that they
were lower class native Britons due to the nomenclature inscribed on the tablets. Bath’s
defixiones also provide a good example of ‘Vulgar Latin’ which also highlights the dedicators
of Bath’s curse tablets. This chapter will outline characteristics of Sulis and why the curse
tablets are important to her cult. The sheer number of curse tablets tells us that they occupied
an important position in Sulis’ cult.
My third chapter explores Sulis as a goddess. This chapter will examine Sulis’ divine
attributes and her role as the patron goddess of Bath. I will discuss the role of religion in
Britain, what religion was like and how this changed under Roman rule. I will discuss the
native British as well as the Roman aspects of Sulis’ cult. This chapter will explain Sulis’
original role as a British deity and how this changed once the Romans took control of her
cult. As Sulis was joined with the Roman Minerva a portion of this chapter will explore
Sulis’ identity as Minerva and evidence concerning Sulis’ role as Minerva. For example, the
gorgon pediment boasts several iconographic features which can be related to Minerva.
Furthermore, most inscriptions identify the goddess as ‘Sulis-Minerva.’ This chapter will
explore evidence at Bath which may reveal potential attributes which Sulis may have
possessed. Bath’s temple was full of architectural features which may illuminate other divine
functions the goddess may have possessed. It is most likely that Sulis’ temple functioned in a
typical Roman way. I will discuss how Sulis’ temple functioned, as well as how rituals were
carried out at the temple. This chapter will outline the layout of Sulis’ temple complex and
how it evolved over time, reflecting the growing popularity of Bath and Sulis’ cult. The
temple’s layout is also suggestive as to how Sulis’ cult would have operated.
10
The goddess Sulis-Minerva was clearly important to the region of Bath as emphasized by her
large temple complex. There is very little scholarship which is solely written as a discussion
of Bath and its patron goddess. Due a lack of primary literature it is very difficult to draw
conclusions about Sulis and her temple. However, Bath is an important example of a large
Romano-Celtic temple complex in Britain. Furthermore, Bath’s defixiones are key examples
of how curse tablets can be diverse. The tablets are significant evidence reflecting the
relationship between mortal and god in Britain. As other parts of the ancient world do not
have defixiones such as these, it is important to further study Britain’s examples. As a
goddess who held such a prominent position, to further study the goddess Sulis-Minerva will
help us to understand religion in ancient Britain more fully.
11
Chapter 1: Bath and Sulis’ Temple Complex
As a site, Bath is difficult to place within a historical context. A large majority of Roman
Bath lies underneath the modern city, making the study of activities at Roman Bath difficult.
There is also very little evidence regarding Bath in ancient literature. Thus, placing Bath
within a chronological context is also difficult. Indicators such pottery sherds and items listed
as stolen on the defixiones provide evidence as to how city life in Bath functioned aside from
the religious. However, there are clues that can be used to outline how the temple of Sulis-
Minerva and the city of Bath functioned when the temple was at its height and began to
collapse. Celtic coinage would suggest that Sulis had been worshipped at Bath since the
British Iron Age.6 Evidence such as temple architecture would suggest that Bath’s popularity
began to climb during the mid-2nd century CE; this popularity reached its peak during the
mid-4th century CE.7 Bath’s prominence as a Roman temple complex was brief in comparison
to others in the ancient world and by the 5th century Bath had ceased to function as a Roman
temple complex. In this chapter, I will outline a historical context for the Roman city of Bath.
Firstly, I discuss Bath’s possible function as a small Roman town and how this is different
from other parts of the ancient world. I outline historical events in Roman Britain that may
have affected the city’s growth and then its fast decline. I also discuss points of physical
evidence that highlight certain points in Bath’s history, such as, the palaeography of Bath’s
defixiones and the presence of pottery and pewter at Bath. It is likely that Bath’s fall was
structured similarly to other Romano-British cities. As a result, I will prove that Bath had an
active temple complex from c. 65 CE-C. 400 CE and prove that although Bath becomes
prominent fast, it’s popularity as a temple complex had a short life span.
6 Green (1986) 155. 7 Cunliffe and Davenport (1985) 179.
12
The History of Aquae Sulis.
If c.175 CE marks the beginning of the deposition of defixiones at Bath, then it is consistent
with the beginning of Bath’s second period and the aggrandizement of the temple complex.8
Bath’s second period is dated to the late second or early third century, corresponding
consistently with Tomlin’s dates for the Old Roman Cursive defixiones in Bath. Importantly,
Bath’s sacred spring became vaulted and sealed during this period.9 The establishment of the
chamber represented a religious shift in Sulis’ temple and an increase in Romanization. The
Romans were generally religiously tolerant towards conquered peoples provided they were
not closed groups or societies.10 Consequently, under Rome ancient Britain was hugely
diverse. For example, Britain retained the majority of its native gods such as Coventina and
Nodens.11 The number of Celtic deities associated with bodies of water was numerous. For
example, the goddess Coventina was also worshipped through a cistern.12 Similarly to that of
Sulis, Coventina’s cistern was the centre of her shrine.13 However, Coventina’s shrine was
open air as opposed to vaulted.14 Goddesses such as Sequana and Sirona also occupied
similar shrines built around sacred springs.15 However, Sulis’ spring was unique in its vaulted
chamber. Although it is likely the inclusion of the vault reflected a change in Sulis’ cult it
may also simply be that Sulis was more popular and therefore had more benefactors. Sulis’
vaulted spring may also have been a reflection of Bath’s growing wealth. Sulis’ spring
received its roof at the same time the temple complex underwent its growth during the
8 Cunliffe and Davenport (1985) 65. 9 For my discussion on restrictions placed on Sulis’ spring see chapter 3. 10 Salway (1981) 665. 11 Mattingly (2006) 215. 12 Coventina’s shrine was not on the same scale as Sulis’ own shrine. 13 Aldhouse-Green (2004) 206. 14 Allason-Jones (1996) 107. 15 Lewis (1966) 48.
13
temple’s second phase.16 Similarly to Sulis, goddesses such as Coventina retain their
identities as Celtic deities.17 However, unlike Sulis they were not equated with Roman deities
such as Minerva. British water goddesses were typically portrayed as nymphs in keeping with
their aquatic associations. Although Coventina, Sequana and Sirona all receive a large
quantity of votive offerings, Sulis is the only ‘water’ goddess who includes defixiones in her
cult.
Furthermore, under Rome Britain developed plenty of purely Classical cults, non-British
cults: an example appears in Chichester with an inscription dedicated to Minerva and
Neptune.18 Bath has similar dedications to Classical gods such as Diana and Mercury.19 Of all
the Classical gods Mercury appeared to be the most popular amongst the Britons.20
Furthermore, eastern gods also became popular in Britain especially towards the second and
third centuries CE.21 The advent of defixiones may represent a shift in how Sulis was
worshipped, evolving from a patron healing spirit of a spring to a goddess who responds to
and answers curses, as the dedication of curse tablets became common practice.22
The second period of Bath’s expansion has been dated to the second century CE. Periods of
little resistance from Britain’s native population helped with the Romanization of Britain.
Romanization would have allowed for Roman buildings such as the Romano-Celtic temple to
be built in Britain’s Roman cities. In addition, Romanization would have helped facilitate the
16 Cunliffe and Davenport (1985) 178. 17 Aldhosue-Green (2004) 206.18 Salway (1981) 666. 19 Although Minerva is worshipped at Bath, she is always worshipped in conflation with Sulis. 20 Green (1986) 36. 21 Salway (1981) 667. 22 Roman religion in Britain will be further discussed in chapter 3.
14
spread of Roman religion easily incorporating Roman religion and culture into Britain’s
southeast region.
By the second century CE, southern England was more settled than its northern counterpart
was. The distribution of Romano-Celtic temples in England shows a large congregation of
temples in Britain’s south (fig. 1). Furthermore, Romano-Celtic temples rarely appear within
Roman Britain’s militarized zone in northern Britain.23 Colin Haselgrove notes that the
distribution of religious centres in England is due to political centralization, with separate
groups of Britons unified under Roman rule.24 Before Roman rule, Britons lived in groups
who were distinct from one another. Within ancient Britain the Celts lived in regional
groupings instead of being unified as a country.25 However, under Roman rule client kingship
was promoted in Britain to further Roman ambitions. Roman client kingship meant that
previously separate groups of people became unified under Roman rule. As South East
Britain was closest to the continent, it was the first to be brought under Roman control and
therefore the first to become politically centralized by Rome. It is important to note that there
is no formal date for a ‘Romanized’ Britain. The process of fusion between Celtic and Roman
religion was gradual and occurred over time. The Celts had been exposed to Mediterranean
culture before the Roman invasion by Caesar in 55 BCE.26 However, the first great upheaval
of British culture came with Claudius’ invasion in 43 CE.
Ancient Britain became more peaceful following the Iceni rebellion led by Boudicca in c.60
CE as the Romans consolidated their territory rather than expanding to the north. After this
23 Mattingly (2011) 225. 24 Haselgrove (2004) 24. 25 Mattingly (2006) 51. 26 Mattingly (2006) 47.
15
period, emphasis was placed on consolidating already conquered land rather than expanding
Roman territory north. The years following Boudicca’s rebellion saw Bath’s first period of
construction in c.69 CE.
David Mattingly notes that regions under British rule tend to display similarities and
continuities.27 Continuity is reflected in Britain’s south-east region with the distribution of
temples. The presence of public amenities such as temples reflects peace in a region as
Romanized towns could be constructed without interruption from native British rebellions.
The second century CE was a relatively peaceful time in Roman Britain. Mattingly notes that
the gaps in primary evidence regarding Roman Britain during this period reflect the lack of
unrest in Roman Britain, noting that problems during Hadrian’s reign were confined to 117-
120 CE.28 Warfare between Britons and Rome during this time was periodic and in short
bursts rather than covering decades. Hadrian’s Wall separated the highly Romanised southern
region of Britain from the wild north. Consequently, there was little development of Roman
temples and public buildings north of Hadrian’s Wall.
Small-inscribed altars seemed to be preferred in England’s northern areas, suggesting that the
army worshipped in ways different to civilians.29 The mobile nature of the Roman Army
meant that it was hard to establish a permanent cult site for worshipping their deities. Small
altars were a quick way to worship one’s god.30 It was also common for the Roman army to
establish portable altars. Although sacrificial altars were commonly permanent, for a group
constantly moving a portable altar would be more convenient. A portable altar may appear in
27 Mattingly (2006) 128. 28 Mattingly (2006) 116. 29 Mattingly (2011) 231. 30 Mattingly (2006) 215.
16
the form of a small metal tripod.31 As there were few settled Romano-British towns in
Northern England, portable altars explain why there is less evidence for Roman religious
practice in England’s northern regions. Britain was slow to urbanize compared to other parts
of the ancient world. Unlike other Roman provinces, urbanism was not well established and
there was no clear administrative system to adapt under Roman rule.32
The height of urbanization occurred in Britain during the second century CE.33 The spread of
temple complexes in Britain matches that of urbanization. Under Roman rule, major towns
fell under three categories: coloniae, municipia and civitates.34 However, although Roman
towns were developing in Britain, their development had started late. Consequently, British
towns never reached the size of other Roman towns such as Pompeii, Herculaneum or
Ostia.35 Coloniae were towns of Roman citizens founded on conquered territory for military
veterans. Municipia were also important towns under Roman rule; ex-magistrates here had
the right to acquire citizenship.36 The civitates were chief towns of a distinct people; these
were defined ethnically and geographically. Although they were of lower status to other
towns, they fulfilled the functions of a local government and stimulated the local economy.37
There were also small towns which were administratively dependent on a colonia,
municipium or civitas centre. Britain’s smaller towns tended to have a more individual
function. For example, a small town may have been a site dedicated exclusively to producing
pottery. Small settlements could also be religious centres.38 In addition, the native elite in
31 Ekroth (2009) 82. 32 Mattingly (2006) 255. 33 Jones (2004) 187. 34 Salway (1981) 574 35 Reece (1980) 78. 36 Salway (1981) 112. 37 Webster (1980) 20. 38 Mattingly (2006) 286.
17
Britain were encouraged to contribute to Britain’s Romanization. Tacitus notes that when
Romanizing the Britons the general Agricola would, ‘adiuvare publice, ut temple fora domos
extruerent’, ‘assist communities, to erect temples, market-places, houses.’ (Tacitus. Agricola
21).
Bath, however, is difficult to place into one of these categories. As Roman Bath lies beneath
modern Bath, discerning Bath’s function in Roman Britain aside from the religious can be
challenging. From what current evidence suggests, Bath was a small town settlement with a
specifically religious function. Mattingly notes that most ‘small towns’ had a very specific
function; this function could be specifically economic or religious.39 From current evidence,
Bath lacked urban aspects that made up larger town complexes, such as the public amenities
of fora, administrative centres, and workshops that typified large Roman towns.40 In typical
Roman urbanism, the forum with appendant buildings constituted the city centre. Flanking
the forum is usually a range of other civic buildings, such as markets, law courts, public baths
and temples.41 Notably, the roads in Roman towns were organized around buildings. Roman
cities were generally aligned on a rectangular grid.42 Bath also had a formal street grid. These
settlements can often have multiple temple complexes. Richard Brilliant notes that Roman
streets were organized through a network of major and minor streets intersecting at rights
angles with quadrangular blocks of insulae, apartments.43
39 Mattingly (2006) 286. 40 Cunliffe and Davenport (1985) 10. 41 Brilliant (1974) 67. 42 Brilliant (1974) 58. 43 Brilliant (1974) 59.
18
Michael Jones states that a typical Romano-British city was organized into settlements of
about fifteen hectares.44 These cities often have a variety of building types housing corporate
activity and industrial production. Typically, a large Romano-British city contained at least
one public amenity. When looking at the distribution of cities in Britain, most large towns
(coloniae or civitates) commonly have two or three public buildings; these are commonly
fora, basilica complexes, theatres and bathhouses (fig. 2). The town of Viroconium
Cornoviorum provides a good example of how a Roman town was laid out (fig. 3).45
Viroconium exemplified the organized town layout with a road system separating the city
into districts, showing the city’s planning. In addition, it has many of the public buildings
which are associated with a Romano-Celtic town. It potentially has two fora with one lying
adjacent to a bathing complex. Furthermore, there are four temple complexes and an
amphitheatre. Similarly, to most Roman towns, Viroconium’s public amenities all lie at the
centre of the city. The forum and bathing complexes lie adjacent to each other at the city
centre. The city was also divided into quadrants similar to those found in more traditional
Roman cities. In addition, the city of Londinium provides a good comparison to Viroconium
(fig. 4).46 Similarly to Viroconium, the city of Londinium was also divided into separate
quadrants with public amenities built near the centre.
Unlike the cities of Viroconium and Londinium, it is likely that Bath fell into the category of
‘smaller towns’ with specific functions. Consequently, Bath’s city plan differed from the
layout of larger cities such as Viroconium and Londinium. For example, Bath had a central
religious focus as a town. Barry Burnham notes that urban growth in Roman towns can be
44 Jones (2004) 164. 45 Jones (2004) 174. 46 Mattingly (2006) 273.
19
distinguished by the presence of a central area.47 Traditionally, the forum and basilica lie at
the centre of a Roman town. However, in small Romano-British towns this does not seem to
be the case.48 Nonetheless, Bath did have public amenities in the form of its public bathing
complexes, the largest of which lay adjacent to the temple complex. Yet, as with many ‘small
towns’ in Britain, the centre of Bath was focused around a religious complex: The Temple of
Sulis-Minerva. Thus, Bath’s primary purpose was mostly likely a religious one. This is clear
even in the city’s Latin name, Aquae Sulis. Bath’s roads all converge upon Sulis’ temple
complex, complementing the temple’s centrality at the site and emphasizing its dominance at
Bath. In terms of size, Bath was significantly smaller than normal Roman settlements.
Furthermore, Bath was surrounded by various civitates and a colonia at Gloucester where
Mercury’s temple is located in the Roman town Uley. Larger towns such as these meant that
supplies could fill the regions need for manufacturers and large markets. Consequently, Bath
was able to develop into a small town developed around a religious precinct. As previously
mentioned, it is unlikely that Bath had a forum. Furthermore, there is no evidence for urban
housing within the site of Bath. Cunliffe notes that due to the architectural scale that a forum
or basilica complex demands, it is unlikely that either were built at Bath.49
The wall around the township of Bath has also been the focal point of discussion. There have
been various theories as to what the wall’s purpose was. There is no specific date given to the
construction of Bath’s outer wall. Cunliffe suggests that the wall was constructed during the
third century CE.50 It is not unusual for ‘small towns’ in Britain to have fortifications, a large
majority of these do so. By the end of the second century CE it was common for both large
47 Burnham, Wacher (1990) 29. 48 Burnham, Wacher (1990) 29. 49 Cunliffe (1971) 66. 50 Cunliffe (1971) 78.
20
and small British towns to include defences surrounding them.51 It can be difficult to discern
why so many of Britain’s Roman towns incorporated surrounding walls. The defence against
hostile forces would have added some incentive to build large walls around towns. For
example, the Boudican revolt saw the destruction of three defenceless towns. The failure to
protect these cities may have encouraged citizens to build more formidable defences around
their cities. Rebellions such as Boudicca’s may help to provide an explanation as to why
Britain built such extensive walls around their cities. Furthermore, the walls would have been
helpful in later revolts. The wall enclosed the majority of the Roman city of Bath with the
temple complex in the centre. As far as archaeological evidence suggests, there were no
extra-mural buildings outside of Bath’s walls.
K. R. Dark notes that Bath may have been a temenos, a piece of sacred land assigned as an
official domain often dedicated to gods.52 Although Sulis’ spring was the most prominent of
sacred springs in Britain, it was not the only one. The idea of sacred bodies of water was just
as prominent in Celtic religion as it was in Roman. Similarly to the ancient Britons, the
Romans believed that natural features such as springs, caves and lakes became the residences
of deities and in turn were revered.53 Typically, these were made into temene (sacred space)
and were entered only for cult practice. The Romans’ treatment of natural temene is mirrored
by the British reverence of natural features. Consequently, British sacred springs were
straightforwardly converted into Roman temple complexes. Bath’s wall marks the territory
which is the sacred area of Sulis. Sulis’ prominence at Bath would lend some weight to
Dark’s hypothesis. Cunliffe also hypothesizes the wall’s potential as a sacred boundary to the
51 Mattingly (2006) 326. 52 Dark (1994) 254. 53 Scheid (2003) 73.
21
gods.54 In sum, archaeological evidence would therefore suggest that Bath was most likely a
religious settlement. As a large religious settlement, Bath’s predominant purpose was the
worship of the goddess Sulis. Consequently, it would be justified to deposit such a large
cache of defixiones.
Dating Bath’s defixiones.
The distribution of defixiones in England is similar to that of Romano-Celtic temples. As of
now, defixiones have been found predominantly in England’s southern region with the
highest percentage found in Bath and Uley. As the dedication of defixiones was a
Mediterranean practice, it is logical that curse tablet dedication would only occur in
England’s south as Roman religious practice may not have been firmly cemented within the
militarized zones. Urban centres, small towns and rural communities have a high rate of
defixiones deposition; however, curse tablets did not seem to be deposited on military sites.55
In contrast, defixiones rarely occur within military communities.56 If British curse tablets
represent a means of communication between a deity and a defigens then a lack of temples
may represent a lack of focal point such as a well or spring. For example, Sulis’ curse tablets
are deposited in the goddess’ spring that is seen as the goddess’ home and best place to
communicate with her. The same can be said of other deities such as Mercury at Uley and the
goddess Sequana. As there were fewer Roman temples in Britain’s northern territories, there
may have been focal points between a deity and mortal. Bath lies well within England’s
Romanized south, which explains the high degree of Roman religious practices. In addition,
British defixiones were all deposited within a temple complex.
54 Cunliffe and Davenport (1985) 10. 55Mattingly (2006) 311.56 Mattingly (2011) 233.
22
The decayed state of Bath’s defixiones makes it difficult to draw any conclusive evidence
regarding dating and content. Roger Tomlin discusses the difficulties associated with
studying and translating Bath’s defixiones.57 Notably, many of Bath’s defixiones were
awkwardly written and semi-legible (Tab. Sul. 6, 16, 40, 75, 77). However, the tablets’
physical condition is the worst impediment to their study. The defixiones would have been
inscribed with a sharp point; however, over time the scratches have faded to the grey colour
of the defixiones making them difficult to read. It is important to remember that the defixiones
were only read once before they were discarded into Sulis’ spring. Importantly, Bath’s
defixiones were made as private inscriptions intended to be read only by the defigens and the
goddess Sulis. Consequently, many of the tablets’ inscriptions were not deeply incised and
wore away over time. Unlike public inscriptions, private inscriptions such as defixiones were
not made to last, and are often unintentionally preserved.58 For example, Bath’s defixiones
were preserved by the collapse of the spring’s reservoir. However, Tomlin suggests that the
majority of Bath’s defixiones can be dated by comparing the incised Latin on the defixiones.59
Bath’s curse tablets are inscribed in a cursive Latin style. The style in which Latin was
inscribed changed over time; this can be an important clue as to when objects that are
difficult to accurately date, such as defixiones, were first created.60 Notably, Francisco Lloris
notes that cursive Latin script with smaller letters was common in private documents.61 Other
examples of cursive Latin can be found on personal letters, poetry and private documents
such as shopping lists. As defixiones were not made to be read by a large public audience,
Lloris’ hypothesis accurately applies to Bath’s own defixiones, particularly when compared to
57 Culiffe (1988) 84. 58 Lloris (2015) 89. 59 Cunliffe (1988) 73. 60 Bruun, Edmondson (2015) 16. 61 Lloris (2015) 89.
23
the large inscriptions on Bath’s publicly dedicated altars. Alternatively to the untidy writing
on the curse tablets, Bath’s altars were inscribed with large capital print making them easy to
read.
In addition, the size of Bath’s defixiones makes translating and perceiving the inscriptions
difficult. The largest defixio, tablet 58, measures 159 by 100 mm (Tab. Sul. 58). But in most
cases, the curse tablets are able to fit comfortably in an adult’s palm. Furthermore, Tomlin’s
dating system excludes the defixiones which do not have an obvious handwriting style as the
lack of handwriting style makes them difficult to decisively date.62 Consequently, 29 of
Bath’s tablets are undated. For example, many of the tablets are written in capital letters.
Tablet 1 one simply states “ABCDEFX” (Tab. Sul. 1) while Tablet 2 states two names
“Britivenda Venibelia” (Tab. Sul. 2). Only a few of the tablets written in capital letters are as
simple as this; many have complete sentences and elaborate formulae.63 For example, tablet 4
states:
qu[i] mihi VILBIAM in[v]/olavit sic liquat com[o](do) aqua. . . qui eam [invol]avit vacat Velvinna ex[s]\upereus Verianus Se/verinus A(u)gustalis Com/itianus Minianus Catus Germanill[a] Iovina.
May he who has stolen VILBIA from me become as liquid as water. . . who has stolen it [or her]. Velvinna, Exsupereus, Verianus, Severinus, A(u)gustalis, Comitianus, Minianus, Catus, Germanilla, Jovina. - Tab. Sul. 4.
Unfortunately, the capital letters on the tablets make them difficult to date. A singular defixio
provides a date stating, ‘ad fontem deae Suli(s) prid(i)e idus Apriles’, ‘at the spring of the
62 Tomlin’s system of dating defixiones is one that has been individually created and is specific to Bath’s defixiones. His dating technique does not reflect one used to date all defixiones. Defixiones also differ in subject and form depending which region they come from. Tomlin’s system is made to date tablets based on the handwriting of Bath’s worshippers, for most of whom Latin was not their first language. 63 For a comprehensive list of Bath’s defixiones written in capital text see Cunliffe (1988) 107- 144.
24
goddess Sulis on the twelfth of April’ (Tab. Sul. 94).64 Unfortunately the tablet does not
provide a year for when it was deposited, highlighting one of the difficulties with studying
Bath’s defixiones.
The 130 defixiones found at Bath can be divided up based on the Latin on the tablets. In
addition to the 29 texts written in capitals, seven of the defixiones are illegible or still rolled
up. Tomlin dates the remaining 94 tablets under Old Roman Cursive (ORC) and New Roman
Cursive (NRC).65 The Old Roman Cursive is dated to c.175-275 CE, placing Bath’s
defixiones deposition roughly a century after the Roman temple complex was erected in c.69
CE. NRC inscriptions were used on the tablets from the fourth century onwards.
Consequently, these represent a much smaller number of Bath’s cache. When comparing
ORC and NRC, there is a clear change in the form of Latin letters when they are inscribed. At
Bath, a transition from ORC to NRC is prominent as there is a clear change in the form of
Latin lettering (fig. 6, 7, 8). The form of the tablets’ Latin provides the only way to place the
defixiones within a context. Comparing the two forms of Latin is useful to discern a rough
timeline of when Bath’s defixiones were being deposited in Bath’s sacred spring.
The first text style is in the ORC style. Tablets 30 through to 93 have been classified as O
RC. Out of Bath’s cache of defixiones, 63 of these are in the ORC style making up the
majority of Bath’s cache. Similarly to Bath, the large number of ORC texts is reflected in
other groups of defixiones and Latin texts.66 Thus we can see a trend in the development of
Latin palaeography. Although the formulae and curses on Bath’s defixiones are diverse, there
64 Unfortunately tablet 94 is the only example of dating on one of Bath’s defixiones. 65 Cunliffe (1988) 73. 66 Cunliffe (1988) 73.
25
is a strong consistency in the inscribed Latin, which is highlighted through the handwriting
inscribed on the tablets since there is a standard format for the way individual letters are
inscribed on the defixiones. Yet, it is also important to note that untidy handwriting should be
taken into consideration when reading Bath’s defixiones. In addition, when discussing ORC
and NRC I am not describing the texts mentioned in capitals as these have no form of cursive
to analyse, although Barry Cunliffe has created referencing tables so we can discern
individual letters on Bath’s defixiones (Fig. 5, 6, 7, 8).67 When referring to the table, the
numbers descending represent which tablet is being represented while the letters to the right
highlight which letters appear on the defixio and what the letters look like. However, these
tables solely represent the Latin on Bath’s defixiones.68
In Bath’s ORC defixiones, the Latin letter ‘a’ is often transcribed in two strokes. On the
legible defixiones the letter ‘a’ is inscribed the same in almost every instance (fig. 6, 7). The
use of two strokes is clear in many examples where the lines do not meet. Furthermore, the
Latin ‘b’ is written in either two strokes or in one loop. In examples from Bath, ‘b’ is always
inscribed in two strokes (fig. 6). Double stroke lettering is evident in most of Bath’s ORC
defixiones. For example, a seemingly simple letter such as ‘c’ is most commonly inscribed in
two strokes. However, certain letters rarely occur in Bath’s texts. The letter ‘k’ only occurs in
a single curse tablet (Tab. Sul. 53); however, ‘k’s were rare in the Roman alphabet as it is a
Greek character. Yet, the letter ‘k’ is used in the word ‘Merkurio’ and ‘Markelinum’ referring
to the god Mercury and a man named Macellinus. Due to the poor state of the defixio and the
semi-literate nature of many of Bath’s defigens, these may be spelling mistakes. Furthermore,
67 Cunliffe (1988) 91-94. 68 See Cunliffe (1988) 91-94 for the tablets.
26
the use of Vulgar Latin is common in Bath’s defixiones. Consequently, the ‘k’ may be an ‘r’
or ‘c’ that was poorly inscribed on the tablet.
The 18 NRC texts are scarce when compared to the large number of ORC texts and of
defixiones written completely in capitals.69 The NRC defixiones can be dated to c. 275-400
CE probably leaning closer to the fourth century due to their small number.70 During the
fourth and fifth centuries CE, NRC texts are particularly rare. Consequently, it can be
difficult to draw consistent trends. NRC writing is distinguished from ORC by its changing
letterforms. Although in some examples of NRC defixiones there are remnants of ORC
lettering which show a transition from ORC to NRC letter forms, the transition is rare at
Bath.71 Examples of letters ‘a’ and ‘m’ provide interesting examples of NRC Latin. The NRC
‘a’ is usually made using one stroke, the ‘a’ forming a closed loop creating an ‘a’ that is close
to its English equivalent. The ‘m’ in NRC is made with one smooth stroke instead of the
multiple strokes which are common in ORC tablets. Some can be identified through unique
formulae. Tablet 66 provides an interesting example of a transition between ORC and NRC.
While the tablet is almost completely inscribed in ORC, the first line is inscribed in NRC
(Tab. Sul. 66). It is likely that Tablet 66 provides an example of a transitional defixiones
probably dating to c.275-300 CE. The author was likely trying a new writing technique
before reverting to one that was familiar.
69 Cunliffe (1988) 86. 70 Cunliffe (1988) 88. 71 Cunliffe (1988) 88. Tomlin notes that the division between ORC and NRC in Bath’s defixiones is very clear. He notes that there was not a large transitional period between the two texts.
27
Despite a change in writing style, the defixiones at Bath continue to ask the goddess to punish
thievery. A contemporary setting can help to identify a NRC curse tablet. For example,
Tablet 98 states ‘Seu gentilis seu Christianus’, ‘Whether pagan or Christian’. The use of
Christianus helps to place the tablet within the fourth century by highlighting the growing
presence of Christianity in Roman Britain.72 In fact, the growth of Christianity in Britain has
been one of the reasons attributed to Bath’s decline from the mid fourth century onwards.73
Importantly, the use of ORC and NRC Latin on curse tablets can provide details as to when
defixiones were predominantly placed at Bath. By comparing Bath’s defixiones with those
found in other Roman provinces, a rough date can be placed defining when written curses
became common at Bath. In terms of ORC Latin, a defixio found at Caerleon provides one of
the earliest examples of British curse tablets (RIB 323, fig. 9).74 The subject matter of the
tablet is the same as that of Bath, revenge for the theft of clothing. Yet, the writing appears to
be a lot earlier than Bath’s own tablets. Comparison with alternative forms of Latin such as
the Dacian wax tablets highlights the fact that Bath’s defixiones are from a much later
period.75 When compared to other contemporary forms of Latin, it can be summarised that
Caerleon’s tablet is likely from the late first century CE or early second century CE.
However, Bath’s ORC examples are dated to around c.175-275 CE placing them later than
Caerleon’s example.76
72 Cleary (2004) 422. 73 Cunliffe and Davenport (1985) 184. 74 Collingwood & Wright (1965) 120. 75 Cunliffe (1988) 88. 76 As of yet there have been no other defixiones excavated from Caerleon. As Caerleon was a Roman fortress it is unlikely many other defixiones will be discovered as curse tablets did not typically get deposited in military locations.
28
Separating Bath’s defixiones into ORC and NRC Latin helps to date the tablet. Although we
cannot date the tablets written in capitals we are able to place a rough date on the remainder.
The majority of the tablets are written in ORC Latin and can be dated to c. 175- 275.
However, the NRC minority are much later most probably dating from the late fourth to the
early fifth century CE. It is important to note that ORC tablets were probably still dedicated
alongside the NRC equivalents. The presence of NRC implies that defixiones may have been
deposited at Bath right up until the fifth century CE.77
The use of metal at Bath.
As a metal, lead is flexible allowing it to be rolled up or folded as displayed in many curse
tablets. This aspect is important as it keeps the curse message as well as the person being
cursed a secret from society’s prying eyes. Lead was both cheap and easily made into sheets
thin enough to inscribe using a stylus or knife, while being more durable than other writing
materials such as papyrus or wood.78 Durability was important, especially if a text was to
survive in a body of water. There are many further advantages to using lead. Lead would
have melted at lower temperatures than bronze and could be made into thin sheets with ease.
Furthermore, lead was also important for sympathetic magic.79 Pliny the Elder remarks on the
availability of lead in Britain noting in his Natural History, ‘Nigro plum ad fistulas
lamnasque utimur, laboriosius in Hispania eruto totasque per Gallias, sed in Brittania
summon terrae corio adeo large, ut lex interdicat ut ne plus certo modo fiat.’ ‘Black lead
which we use to make pipes and sheets is excavated with considerable labour in Spain and
77 Cunliffe (1988) 97. 78 Cunliffe (1988) 81. 79 I discuss this further in chapter 2. Within Bath’s cache only tablet 4 has an example of sympathetic magic stating that the victim should ‘become as liquid as water’ (Tab. Sul. 4). Sympathetic magic may not have been regularly practised at Bath.
29
through the whole of the Gallic provinces, but in Britain it is found in the surface-stratum of
the earth in such abundance that there is a law prohibiting the production of more than a
certain amount’ (Pliny. Nat. 34.49). According to Pliny, lead was plentiful in Roman Britain
and much easier to obtain than in other regions of the Roman Empire.
However, the majority of Bath’s defixiones were not made from a pure lead alloy. Many of
Bath’s curse tablets were diluted with tin or copper alloys. As Cunliffe notes, most of Bath’s
defixiones are closer to pewter than lead.80 Unfortunately other British tablets have not been
analysed and we are unable to discern if Bath was unique in the use of pewter defixiones.
There is no evidence to suggest that tin, copper or pewter had any magical properties that
could be used with sympathetic magic.81 It is more likely that Bath’s tablets were not made of
standardized lead for pragmatic reasons. For example, pewter was more readily available in
the ancient world, more durable and importantly, cheaper than lead. Michael Fulford
meanwhile discusses the evidence of silver, lead, copper and tin mining in Britain. Stamped
ingots provide the primary evidence to these activities in Britain.82 Silver was regularly sent
back to Rome while cheaper metals such as lead and copper were consumed within the
province.83 Lead and pewter were readily available within Britain for creating things such as
defixiones and the water pipes for Bath’s large bathing complex.
In addition, pewter was commonly used in many of Bath’s metal vessels, notably inscribed
paterae. Similarly to Bath’s defixiones, many of these were inscribed, often as dedications to
80 Cunliffe (1988) 82. 81 Unfortunately there is not really any primary evidence regarding the use of sympathetic magic in the ancient world. 82 Fulford (2003) 317. 83 Fulford (2003) 318.
30
Sulis. Currently there are 23 vessels recovered from Bath, 20 of them being made from
pewter.84 These vessels were plates (RIB 106), dishes (RIB 99), jugs (RB 100), and most
commonly pans (fig. 10).85 The pewter pans found at Bath were different from Roman
paterae which were commonly used for libations. These pans were much deeper than that of
a typical Roman patera, and these pewter pans would have been able to store more liquid in
them. There are shallower pewter ‘plates’ amongst Bath’s vessel hoard, but, when compared
to the pans, the plates are plain in decoration and carry no inscriptions. Interestingly, Tomlin
describes Bath’s pewter pans as ‘paterae’ in his discussion of the vessels’ inscriptions.86
Seven of the pans have inscriptions. All of the inscriptions read ‘Deae Suli Minervae’ ‘To the
goddess Sulis Minerva’ (Britannia xii 23) or ‘Deae Suli’ ‘To the goddess Sulis’(Britannia xii
19). The inscriptions firmly identify the paterae as votive offerings to Sulis. Furthermore, the
inscribed paterae are often accompanied by decoration. When compared to the plainer pewter
vessels, Tomlin may be justified in his assumption that these are paterae.
While many of the vessels can be considered ‘household objects’, the distribution of pewter
vessels likely represents ‘votive’ deposition rather than actual pewter use as they were found
in Sulis’ spring.87 The distribution of pewter in Britain was very uneven.88 For example,
Roman towns that provide evidence of vessel manufacture may yield few finds. Pewter
vessels tended to appear further away from cities where vessels were created or mining
occurred. Romano-British pewter manufacturing towns are traceable through the waste left
behind by manufacturers. For example, the distribution of moulds, tin alloy and pewter scrap
84 Cunliffe (1988) 9. Three of the recovered vessels are made of bronze or silver. Two of them have inscriptions that identify them as votives for Sulis. 85 Cunliffe (1988) 55. 86 In their descriptions and illustrations of the pewter vessels Nigel Sunter and David Brown refer to Tomlin’s paterae as ‘pans’. 87 Beagrie (1989) 191. 88 Hatcher & Barker (1974) 10.
31
alludes to the creation of pewter vessels at a site.89 A map created by Neil Beagrie highlights
the distribution of pewter manufacturing scrap wares in Roman Britain (fig. 11).90 Beagrie’s
map notes that the highest amount of pewter manufacturing scrap was found in areas
surrounding modern Bristol. Yet, when comparing a map of manufacturing finds to
distribution of pewter vessels, the largest caches of pewter vessels are found between modern
London and Cambridge. Pewter vessels have been most commonly found in wells, rivers and
ponds, following a common practice of votive deposition in England which is clearly also the
case at Bath. However, unlike other Romano-British pewter deposits, Bath is a short distance
from Bristol where moulds, tin alloy and pewter scraps were found. Furthermore, the
distribution of pewter manufacture is also important. There was a large concentration of
pewter around the Bath area.91 For the mass production of curse tablets, a cheap and easily
malleable metal would have made the process of creating curse tablets much faster.
The rich pewter content of Bath’s votive offerings can help place a rough date on Bath’s
defixiones. In his analysis of the Romano-British pewter industry, Beagrie notes that the
majority of Romano-British pewter was retrieved within a late Roman context with a
proposed date of c.250- 410 CE.92 John Hatcher’s consideration of British pewter also notes
that British pewter was manufactured between c.250- 420 CE.93 Interestingly, the dates for
British pewter manufacture are similar to the dates Cunliffe suggests for Bath’s defixiones.
Similarly to Roman votive offerings, the majority of pewter is deposited in England’s
southern region.
89 Beagrie (1989) 178. 90Beagrie(1989)177.91 Beagrie (1989) 178. 92 Beagrie (1989) 175. 93 Barker, Hatcher (1974) 10.
32
The tablets from Vindolanda provide a good example of ORC Latin in Britain and can be
compared to the handwriting on Bath’s own tablets. Vindolanda’s tablets were made out of
wood with a thin covering of wax. Over time the wax eroded leaving only the wood,
however, the writers inscribed with enough force to leave marks on the wood.94 The
defixiones and Vindolanda tablets are different in form and purpose, however, the act of
inscribing is similar in both defixiones and wax tablets. Both the Vindolanda tablets and
Bath’s defixiones are invaluable as examples of ORC and its development. However, the
Vindolanda tablets are dated to c.90-120 CE meaning they were made before Bath’s curse
tablets.95 Through the use of lead and tin, however, Bath’s defixiones conform to the
traditional material used for creating curse tablets.96
In contrast to those at Bath, the Vindolanda tablets are both personal and administrative
documents.97 Additionally, the Vindolanda writing tablets are dated much earlier than Bath’s
defixiones. Alan K. Bowman notes that the Vindolanda tablets were created shortly after c. 90
CE.98 The dates are important as, by the dates proposed for Bath’s defixiones, c. late 2nd-5th
century CE, Latin palaeography had changed. Bowman and Thomas mention Bath’s
defixiones in their consideration of Vindolanda ORC; they concede that Bath’s inscribed
Latin is different from that of Vindolanda’s, noting that the two make a poor palaeographic
comparison.99 Furthermore, the techniques used for inscribing lead and writing on wood
would have slightly differed, creating different letter forms. finally, the Vindolanda tablets
94 Bowman (1994) 8. 95 Bowman (1994) 10. 96 Graf (1997) 133. 97 Bowman, Thomas (1994) 18. 98 Bowman (1994b) 6. 99 Bowman, Thomas (1994) 48.
33
were inscribed by Roman soldiers and officials.100 Consequently, the handwriting is far more
practised than the semi-legible inscriptions on Bath’s defixiones.
The differences between the Latin inscribed on Bath’s defixiones and the script on the
Vindolanda tablets highlights the changing ORC script (fig 12). Bath’s curse tablets have
letters that deviate in form from the Vindolanda tablets. For example the use of the letter ‘n’
differs greatly between the two sets of texts. While the ‘n’ in the Vindolanda tablets
resembles the modern day ‘n’ in form (fig. 12), the Latin ‘n’ on Bath’s defixiones resembles a
reverse y-shape (fig. 6, 7). Furthermore, the use of the letter ‘e’ in Bath is different to that of
earlier ORC Latin (fig. 6, 12).101
When Roman Bath ended.
It is notable that Bath seemed to have reached its high point by the middle of the 3rd century
CE. The only ancient text to mention Bath is Solinus’ Collectanea rerum memorabilium.
Unfortunately Solinus’ account of Bath is very brief stating, ‘quibus fontibus praesul est
Minervuae numen, in cuius aede perpetui ignes numquam canescunt in fauillus, sed ignis
tabuit uertit in globos saxeos.’ ‘over these springs Minerva presides and in her temple the
perpetual fire never whitens into ash but as the flames fades turns into rocky lumps’ (Solinus.
De Mirabilibus Mundi XXII). Solinus confirms that Sulis was regarded the primary deity at
Bath and the patron goddess of the hot springs, at least in Roman eyes. The perpetual fire is
also an interesting statement and alludes to constant vigilance over Sulis’ temple complex.
100 Bowman (1983) 27. 101 Tomlin (1988) 87.
34
Notably, when Solinus visited Bath during the third century CE the temple was fully
functional and prevalent.
While the third century CE saw Bath’s height as a religious site due to peaceful times,
Britain’s fifth century is one that is full of discord. By the fifth century CE Roman rule in
Britain had passed its peak and entered a precarious period leading towards an era in Britain
without Roman guidance. Ian Wood notes that the beginning of the end of Roman rule can be
linked to the revolts in Britain which concluded with the elevation of Constantine III as
emperor in c.407 CE.102 However, from 410 CE the Britons existed without any Roman
authority. With receding Roman interest in Britain, it is likely that the maintenance of public
buildings in Ancient Britain was in short supply. Despite the absence of Roman control,
British towns continued to function after the Romans left.103
Cunliffe suggests that the decline of Roman Bath was prolonged and uneven depending on
which parts of the town were being affected.104 Regardless, by the middle of the fifth century
Bath’s temple and bathing were abandoned and no longer in use.105 The popular opinion held
is that the prominence of Christianity meant that traditional Greco-Roman religion was
phased out. Bath did not appear to be immediately converted to Christian purposes.106 In
Bath’s case many other factors such as political disunity, economic decline and Celtic
rebellion led to the temple’s decline in maintenance and eventual abandonment. Bath fits into
larger patterns throughout Roman Britain portraying the decline of Roman culture and
102 Wood (2004) 432. 103 Gerrard (2007) 149. 104 Cunliffe (1971) 91. 105 Cleary (1989) 131. 106 Wood (2004) 428.
35
influence. An eighth century Anglo-Saxon poem, The Ruin, describes the state of Bath
during this period.107 The Ruin notes, ‘Wondrous is this masonry, shattered by the Fates. The
fortifications have given way, the buildings raised by giants are crumbling. The roofs have
collapsed; the towers are in ruin….. There is rime on the mortar.’108 The poet describes
Roman ruins that have fallen into disrepair, a state of affairs which was not uncommon in
Anglo-Saxon Britain. The poem does not specifically mention Bath by name, though it does
mention ‘The wall enfolded within its bright bosom the whole place which contained the hot
flood of baths’: Bath is the only known Roman bathing complex built around thermal springs
in England. Interestingly, Cunliffe notes that The Ruin matches his hypothesis on how Bath’s
temple and bathing complexes fell into disrepair.109 Presumably defixiones stopped being
deposited after this period. Within large towns, public amenities were being abandoned and
disused. Furthermore, deterioration in the maintenance of roads and drains highlights the
neglect of Roman cities.
In terms of religion, pagan cults reached their peak in the 4th century CE. After 300 CE
Roman religious complexes followed the same pattern as public buildings and baths. Over the
course of the 4th century, there was a decrease in the number of temples in Roman Britain.110
Many scholars such as David Mattingly and Richard Reese suggest that the Romanized town
system had failed as early as the third century CE.111 However, this is not suggesting that
towns in ancient Britain had collapsed by this period but rather that they did not follow a
strictly Roman form. Mattingly argues that life in towns became more related to ‘individual
107 Unfortunately the author behind the The Ruin is unknown. There is also no specific date that the text was published. 108 Rime is an ice that forms on objects through rain and fog. 109 Cunliffe (1971) 95. 110 Cleary (1989) 119. 111 Mattingly (2006) 325, Reece (1980) 77.
36
power’ than ‘civil authority’.112 One of the key pieces of evidence for Mattingly’s hypothesis
is the state of public buildings in British towns. For example, the forum and basilica complex
in Silchester was given over to metalworking by the end of the third century.113 Similarly, the
basilica complex in Caerwant was demolished and replaced with a metal workshop similar to
that in Silchester.114 In other cases, public buildings in towns such as Wroxeter and Leicester
were simply not replaced after they had fallen.115 In ancient Britain there is a notable loss in
civic spaces, but Roman public buildings were reused rather than abandoned, emphasizing
the continuation of urban British life.
The reason for the dilapidation of public buildings in part stems from a lack of funding from
the local elite. Interestingly, British urban elite continued to follow a Roman pattern of design
and decoration.116 The furnished state of personal homes is at odds with the continuing disuse
of civic amenities. It seems, therefore, that the urban elite were funding personal expenditures
rather than public. Reece’s hypothesis regarding the collapse of the Roman-British town is
similar to Mattingly’s. Reece suggests that a revival of town life was impossible due to the
establishment of a provincial structure based on villas and villages.117 In terms of religion,
this meant that large temple complexes did not receive the funding they needed to maintain
their grandeur.
Ken Dark discusses the role of the elite in the survival of pagan religion during late antiquity.
Dark notes that there was a pagan religious elite while Christianity flourished within Britain’s
112 Mattingly (2006) 325. 113 Mattingly (2006) 336. 114 Mattingly (2006) 336. 115 Mattingly (2006) 326. 116 Mattingly (2006) 339. 117 Reece (1980) 78.
37
lower class during the fourth century CE.118 However, Dark’s hypothesis does not seem to
apply in the case of Bath. Although sites such as Bath clearly received attention from elite
classes and military personnel, there was also a large amount of activity from the lower
classes as well. When evaluating defixiones, it is clear that they were mainly deposited by the
lower class, often British citizens, as is clear at temple sites such as Uley and Bath.
Sulis’ temple complex fell into disuse around the same time as other temple complexes in
Britain. Chronologically it is difficult to place a fixed date on Bath’s end as a Roman site.
Nonetheless, Bath’s temple precinct, Sulis’ temple complex, began shrinking from c.350 CE
when the focus was restrained to the inner precinct of the temple.119 For example, the
stylobate leading toward the temple complex’s altar courtyard was dismantled and the
colonnade surrounding the temple and courtyard was also dismantled giving the complex a
far more secular appearance. The decaying state of Sulis’ religion is further emphasized by
the dismantling of the great outdoor altar, which represents one of the temple’s most
important religious monuments.120
In addition, layers of cobbling and soil can tell us about Bath’s dilapidation. From the late
fourth century onwards the temple was built on a series of superimposed floors as soil began
to accumulate in Bath’s inner precinct.121 Although silting was occurring, Bath’s temple
precinct remained in use and cobbling was used to create pathways. Eventually soil gathered
on the cobbling and once too much accumulated a fresh layer of cobbling was added to create
118 Dark (1994) 38. 119 Cunliffe and Davenport (1985) 184. 120 Gerrard (2007) 159. 121 Cunliffe attributes this to the blockage of gutters and drains.
38
a new pathway. Within these layers rubble, debris and pottery sherds were trapped creating
useful pieces of evidence for dating Bath’s final phase as a Roman temple. According to
Cunliffe, the constant reflooring raised the level of the precinct by nearly a metre higher than
the original paving.122 Importantly, this meant that many monuments and altars in Bath’s
courtyard were slowly buried.
In terms of religion, the continued decay of the temple’s buildings and facilities is seen in
sedimentary levels and the material used to remake pathways and floors.123 In many cases,
Bath’s buildings were deconstructed to be used in the new paving. Cunliffe reports that a
piece of the temple’s column was found in the pathing.124 It is clear that in Bath’s later period
repurposed building material was used to create the cobbling which maintained the flooring
and paving. Importantly, Bath’s religious buildings were being deliberately demolished.125
The demolition of the rest of the temple complex meant that thieves could gain access to the
lead and iron clamps which kept the blocks in place. Furthermore, the buildings may have
been demolished to reuse the metal and stone. The intentional demolition of such a
previously important religious site highlights the waning influence of Greco-Roman religion
in later fourth century CE Britain.
Evidence suggests that Sulis received offerings up until the fifth century CE. It is useful to
return to the Anglo-Saxon poem The Ruin. To support this Cunliffe suggests that Sulis’
temple complex became the victim of flooding from the spring. In a way Bath collapsed from
122 Cunliffe and Davenport (1985) 72. 123 Cunliffe and Davenport (1985) 169. 124 Cunliffe and Davenport (1985) 185. 125 Gerrard (2007) 159.
39
the inside out. If Cunliffe’s hypothesis is true, then Bath’s religious and recreational
functions, arguably the city’s primary features, were the first to fall into disrepair. However,
there are ways in which we can discern the years that Bath was a functional township. One
of the most convenient ways to place Bath into a chronology is through its pottery. In their
consideration, Sarah Green and Christopher Young note that the significance in pottery from
Bath lies in its lateness.126 Fortunately, Bath was a well-maintained site up until the fourth
century CE. Consequently, there is little evidence pottery-wise from before the fourth
century.
The majority of pottery excavated from Bath is dated to its final period of Roman usage.
Fortunately, the sedimentary layers of Bath’s final period mean that the pottery sherds are
relatively well preserved. Accordingly, we know that the temple complex continued to
function at least in part despite the decline in the second half of the fourth century.
Unfortunately, however, much of the pottery found at Bath is in poor condition; this makes
discerning any figural decoration difficult. The pottery recovered from Bath was almost
exclusively made in a local British style (although influenced by Mediterranean examples in
form and purpose).127 Notably, from the third century CE onwards pottery was made solely
in England rather than imported from the continent; this is reflected in a lack of Samian
ware.128 A large number of Bath’s vessels include fabrics from recognised Romano-British
pottery industries such as those in Oxford, New Forest and Alice Holt.129 The kilns for each
of these types were close in proximity to Bath so transport would not have been difficult.
126 Green, Young (1985) 157. 127 Taken from my own tallies of the fabrics used in Bath’s pottery. My tallies were taken from date found in Cunliffe, B. (1985) The Temple of Sulis Minerva at Bath Volume 1: The Site. Oxford: Oxford Committee for Archaeology, p 143-161. 128 Fulford (1973) 164. 129 Cunliffe and Davenport (1985) 146.
40
Thirteen examples of the late pottery recovered from Bath can be attributed to kilns from
Oxford; there are also seven New Forest vessels and six vessels from Nene Valley.130
However, as time progressed both the Nene Valley and New Forest kilns were displaced by
Oxford workshops. The same trend is apparent in Bath; at the beginning of the middle fourth
century CE Oxford and New Forest ware were found in relatively similar amounts. However,
by the beginning of the fifth century CE only Oxford examples were being discovered
alongside the local ware.
The vast majority of Bath’s pottery was made from local fabrics and kilns. Because the
modern township of Bath is inconveniently built on top of the Roman city, it is hard to find
evidence that reflects economic activity at Bath. Mattingly suggests that it is not uncommon
to find marketing activities in religious sanctuaries.131 A prominent sanctuary like Bath would
have likely included some form of economic activity; the presence of local pottery highlights
this. Bath’s pottery consisted of an assortment of pottery types including jugs, jars, beakers,
flagons, mortaria and bowls.132 As far as we are able to tell, none of Bath’s pottery boasted
inscriptions which would identify them as votive offerings. It is most likely that they were
manufactured for storage and carrying goods. However, some of the pottery was painted with
simple designs such as rosettes and geometric designs. Paul Tyers suggests that Bath local
ware fits into a distribution pattern where a local kiln produces a particular pot which is
widely distributed; the form and fabric is then copied by local workshops which then mass-
produce imitations.133 Such imitation pottery is reflected at Bath through the reduction of
wares from notable kilns such as New Forest. Over the third and fourth centuries CE, pottery
130 Cunliffe and Davenport (1985) 146.131 Mattingly (2006) 290. 132 Cunliffe and Davenport (1985) 153. 133 Tyers (1976) 77.
41
from foreign workshops decreased at Bath. However, the production of local wares continued
to rise until the late 5th century. The economic decline in England during the fourth century
may have prompted a decline in sources of pottery supply. Consequently, local ware
imitations were made to replace the absence of notable pottery kilns. From the middle of the
fourth CE, pottery was in decline with a reduction in the number of vessel shapes and a
decline in decorative techniques.134 For example, fine ware pottery gave way to strictly
coarse ware.
The pottery from Bath’s late period is useful for creating a context for Bath’s history. Pottery
continued to be used at Bath right up until the end of the Roman Period. Interestingly,
although Bath’s religious complex was going through a rapid decline, pottery continued to be
used at the site. The decline in pottery is important when highlighting Bath’s decline: pottery
from large kilns declined while plainer local wares were created in their place. Consequently,
we can identify that despite the temple’s decay, some sort of activity continued at the temple
complex. However, we cannot identify how much of this is religious as Bath’s religious
decline was clear through the temple’s dereliction.
In addition, over the span of its Roman use, the site became a repository for a large collection
of coinage. Most of the coinage was found in the sacred spring pointing to the use of coins as
a votive offering to Sulis.135 The decoration of coinage can provide useful evidence
pertaining to the date when the coins were dedicated. Furthermore, coinage helps to identify
what periods saw the heaviest amount of activity at Bath as well as the most offerings.
Although we cannot say that defixiones were dedicated on the exact same dates, coinage may
134 Tyers (1996) 76. 135 Cunliffe (1988) 281.
42
help to outline a trend in dedications at Sulis’ spring. The periods with the largest amount of
coinage were likely Bath’s busiest time. Coin deposits at Bath have been dated back to Bath’s
Pre-Roman era suggesting that coin votives had been a common form of offering to Sulis.136
The coins themselves date to c.10-40 CE when the British king Cunobelin reigned.
Unfortunately, like many artefacts recovered from the sacred spring, the coins have been
badly corroded by Bath’s spring water. Furthermore, the Celtic coins are the earliest offerings
found in the city of Bath. It is clear, though, that coinage from after c.260 CE was discovered
in huge quantities.137 The large amount of coinage fits within the period where Bath’ temple
complex saw a high amount of activity. Consequently, the third century onwards seemed to
be the period when Sulis was receiving the highest amount of offerings.
The practical lifespan of Bath’s own temple complex ended with the collapse of the spring
reservoir, which effectively ended the deposition of defixiones in Bath’s spring complex.
Through the demolition of the temple complex and the dereliction of the site, Bath fell into
disrepair. Furthermore, the dereliction of Roman temple complexes coincides with the date
attributed to the NRC tablets. Consequently, this may be why Bath has so few NRC tablets in
comparison to ORC defixiones. The NRC tablets may represent the defixiones from Bath’s
final days.
One of the hypotheses associated with the end of Roman Britain was the ‘native revolts’
which took place c.340-410 CE.138 From c.360-400 CE, Picts and Scots began raiding the
136 This is a very small deposit when compared to the large hoard of Roman coinage which was discovered at Bath. 137 Walker (1988) 281. 138 Jones (1996) 244.
43
frontier line causing destabilization in the province of Briton. Thus, the backdrop for political
discord was formed. It is likely that this destabilization occurred with the rising of three
usurpers Emperors, Marcus, Gratian, and Constantine III. Michael Jones notes that barbarian
revolts played a significant role in the end of Roman Britain.139 As Gaul began to revolt
Constantine III drew forces from Britain to settle the Gauls as well as to conquer Italy.
Although Constantine succeeded in Gaul, he failed to take Italy. Jones notes that due to the
lack of troops, and the troops’ lack of faith in their leaders, the early 5th century CE saw the
end of Roman authority in Britain.140 Through Roman political conflict the Britons were able
to remove the yoke of Roman authority.
Notably, primary sources reflect the Roman abandonment of imperial Britain. In his Historia
Brittonum Nennius notes,
‘Hucusque reganverunt Romani apud Brittones ccccviiii annis. Brittones autem deiecerunt regnum Romanorum neque censum dederunt illis neque reges illorum acceperunt, ut regnarent super eos, neque Romani ausi sunt, ut venirent Brittanniam ad regnandum amplius, quia duces illorum Brittones occiderant.’
‘Hitherto the Romans had ruled the British for 409 years. But the British overthrew the rule of the Romans, and paid them no taxes, and did not accept their kings to reign over them, and the Romans did not dare to come to Britain to rule anymore, as the British had killed their generals.’ (Nennius. Historia Brittonum 28).
The constant raids and destabilization had created an opening to reject Roman rule in Britain.
As a consequence there was a rejection of the previous romanization of Britain. Without
Roman rule Romano-British temple complexes began to collapse. Romanized cults such as
Bath had ceased to function. From the early 5th century onwards Britain was no longer a
Roman province.
139 Jones (1996) 246. 140 Jones (1996) 248.
44
Conclusion.
To conclude, Bath’s temple complex is very difficult to date. Many of the hypotheses
surrounding the town of Bath, the temple complex and the cult of Sulis-Minerva are made by
comparing Bath to other Romano-British towns and temples as well as examining the
physical evidence at Bath. Bath fitted into a small group of Roman towns that often had very
specific functions with the town built around an economic or religious building. Due to the
size of Bath, many of the typical features of a Roman city do not exist at Bath. For example,
Bath lacks Roman public amenities such as a forum. The wall around the city Bath likely
acted as a temenos outlining the sacred boundary of Sulis’ precinct. As a continuation of its
Celtic roots, Bath most probably existed as a religious town with its focus on Sulis’ cult. It is
likely that Bath’s period of most prominence started in the 3rd century CE and lasted through
until the 5th century. The palaeography on Bath’s defixiones helps to place Bath within the
previously mentioned chronology. The Old Roman Cursive is fitting for private Latin texts
from the ancient world, while the New Roman Cursive tablets were most likely from the
temple’s last days. The far smaller amount of New Roman Cursive attests to the collapse of
the temple. Furthermore, coin hoards found in Bath’s spring help to highlight Bath’s periods
of use. Through the destabilization of Roman Britain and changing religious sensibilities
Bath was abandoned as a Roman temple complex. However, Bath continued to be inhabited
right up until the modern era. Although portions of Bath, predominantly the temple and
bathing complexes, have been uncovered a large portion of the city still remains buried
underneath Bath’s modern city. Consequently, there is still much to learn about the Roman
town of Bath.
45
Chapter Two: Prayers or curses? A study of Bath’s defixiones.
Bath’s defixiones have aspects which contrast with what are considered ‘traditional’ curse
tablets. Because the defixiones were written on tablets and then deposited in a spring, it is
easy to consider Bath’s tablets as a form of defixiones. However, on closer inspection, there
are several elements which set defixiones such as Bath’s apart from ‘typical’ curse tablets. In
his consideration of defixiones, H. Versnel suggests that several tablets should not be
regarded as curses but instead prayers for justice. Although they resemble defixiones, tablets
such as Bath’s should form a new category altogether. Versnel places Bath’s defixiones under
the category of ‘judicial prayers’ or ‘prayers for legal help.’141 When regarding Bath’s
defixiones, one can see why the label ‘prayer for help’ is appropriate. Solinus asks Sulis for
her aid stating:
Deae Suli Minervae Solinus dono numina tuo maniestati paxsam balnearem et paleum nec permittas somnum nec sanitatem <.> ei qui mihi fraudem fecit si vir si femina si servus si liber nissi [<s>s]e retegens istas species ad templum tuum detulerit li[beri sui vel son sua et qui.142 deg. ei quoque xe. so,mi, nec sanitatem n al<u>leum et reli<n>quas nissi ad templum tuum istas res retulerint.
Solinus to the goddess Sulis Minerva. I give to your divinity (and) majesty (my) bathing tunic and cloak. Do not allow sleep or health to him who has done me wrong, whether man or woman, whether slave or free, unless he reveals himself and brings those goods to your temple. - Tablet. Sulis. 32.
Solinus’ prayer to Sulis is not only an appeal for the goddess’ help, but also a request for
justice as he feels he has been wronged. Therefore, Versnel’s consideration is well grounded
as Bath’s defixiones are different in many aspects to other tablets.
Solinus’ plea is not the only prayer from Bath to address the theft of personal belongings. In
141 Versnel lists 7 points which help to highlight how ‘judicial prayers’ are different from defixiones. See Magika Hiera (1991) 68. 142 Here the tablet is badly damaged. Holes in the defixiones make translating them difficult.
46
fact, the majority of Bath’s defixiones deal with cases of theft. A victim who has had his
property stolen asks for the goddess’ help in the retrieval of his (or her) stolen property.
However, if the stolen goods cannot be returned, an alternative form of justice is the
punishment of the thief. As previously noted, these ‘prayers’ were placed in Sulis’ spring so
as to communicate directly with the goddess. Interestingly, there is not a single piece of
evidence to suggest Bath’s tablets dealt with another form of cursing other than judicial.
However, Bath’s judicial prayers are not a phenomenon that is unique to Bath. Notably, a
similarly large cache of tablets found in Uley, England, provide a good comparison to Bath’s
own defixiones. Like Bath’s, Uley’s defixiones ask a deity (Mercury) to recover stolen goods
from a thief. A tablet found in a temple dedicated to Mercury provides a useful comparison
to Bath’s own tablets:
Side A: Deo Mercurio Cenacus queritur de Vitalino et Natalino filio ipsius d[eiumento quod eraptum est. Erogat deum Mercurium ut nec ante sanitatem Side B: habeant ni<s>i ni<s>I repraese[ntaverint mihi iumentum quod r[apuerunt et deo devotionem qua[m ipse ab his expostulaverit.
Side A: Cenacus complains to the god Mercury about Vitalinus and Natalinus his son concerning the draught animal that was stolen. He begs the god Mercury that they will not have good health…
Side B: …until they repay me promptly the animal they have stolen and (repay) the god the devotion which he himself has demanded from them.143
Although not identical to Bath’s defixiones, Uley’s defixiones also address cases of theft:
Mercury is asked to hunt down the thief and force compensation from him. Notably, the
majority of defixiones found in Britain all fit into a category of curses against thievery.144
Barry Cunliffe notes that there is not a single instance of a curse which is directed against
charioteers or lawsuits in British curse tablets. Curses against theft, which were so common
143 Britannia 10 (1979) p. 342, no. 2. 144 Cunliffe (1988) 60.
47
in Britain, were not the predominant topic for curses in the rest of the ancient world. Lindsay
Watson notes that revenge curses can be used to appropriate the functions of the law code.145
If one felt that the law could not provide satisfaction to the wronged, it was instead possible
to turn to the gods for help. As I will explain later in this thesis, those most likely to turn to
Sulis for legal representation seemed to be Bath’s lower socio-economic classes; this is
revealed by the names of Bath’s defigens as well as the Latin inscribed on the tablets. While
it is correct to read Bath’s as ‘prayers for justice’, they should still be considered as a
category of defixiones. Bath’s tablets still conform to several characteristics that are common
of defixiones such as where they were deposited and the punishments inscribed on the tablets.
This chapter will prove that Bath’s curse tablets are an example of a British adaptation of
Greco-Roman defixiones.
Vulgar Latin on Bath’s defixiones.
Bath’s tablets are an important source of Vulgar Latin from a Roman town in Britain. Jozef
Herman defines the term ‘Vulgar Latin’ as ‘the spoken language of people who were scarcely
influenced at all by literary tradition’.146 He refers to a form of Latin which is common and
‘colloquial’. Generally, Vulgar Latin is inscribed the way it would be spoken in a
conversation. Vulgar Latin differs from literary Latin in many ways, although for brevity’s
sake I will only mention differences that are relevant to Bath’s defixiones, which were not
written as high literature. Furthermore, the evidence available in Britain pertaining to written
Latin is fragmentary, making it difficult to distinguish a pattern in British Latin. The range of
available epigraphic evidence from Roman Britain has greatly increased over the with the
discovery of a cache of letters inscribed on wood from Vindolanda. The Vindolanda tablets
145 Watson (1991) 6. 146 Herman (2000) 7.
48
provide a source of evidence for colloquial Latin written in Britain.147 These letters differ
from Bath’s tablets as they are not intended as religious pieces of writing. However, they
provide further examples of Latin which was not written as Classical or later literary Latin.
As Bath was a provincial town on the fringes of Rome’s Empire, it is expected that the Latin
used would vary from that used in Rome. Colin Smith’s consideration of British Latin is
important when discussing Bath’s defixiones. Smith notes ‘that even uncultured writers of
graffiti and defixiones were conscious of the learned and classical tradition means that we
have no real Vulgar Latin texts for Britain.’148 Therefore, we should not regard the Latin
inscribed on Bath’s tablets as incorrect but rather as ‘intrusions’ onto the Classical Latin.
Although vulgar terms may be spelt differently from Classical Latin on British tablets, the
terms fit correctly into a sentence. Due to lack of literary evidence in many cases, it can be
difficult to discern what constitutes ‘typical’ Latin in Roman Britain. Although Bath’s
defixiones share many common features with other tablets (especially British) it can be very
difficult to know if the terminology or spelling is typical in Britain. All we can say for certain
is that many features of Bath’s defixiones do differ from Classical Latin.
When discussing the intrusion of Vulgar Latin into Classical Latin it is important to
remember the highly formulaic nature of the defixiones. Language in Bath’s tablets often
includes variations of similar phrases.149 For example, one of the most common phrases
found on Bath’s defixiones is, ‘devoveo eum [q]ui caracellam meam involaverit si vir si
femina si servus si liber’ ‘I curse him who has stolen my hooded cloak, whether man or
woman, whether slave or free’ (Tab. Sul. 10). The catch-all phrase on tablet 10 is a feature
147 Bowman (1994) 1. 148 Smith (1983) 936.149 Cunliffe (1988) 63.
49
that is prevalent in many of Bath’s defixiones, an aspect which may have been used to help
cast a wide net over potential victims, especially if the dedicator did not know the identity of
the thief. Examples of tablet 10’s catch-all formula can be found on several of Bath’s
defixiones emphasizing the heavy use of catch-all formulae on Bath’s curse tablets (Tab. Sul.
32, 36, 49, 52, 71, 100). So, we can see how this aspect of cursing was highly formulaic due
to the regularity in which the phrase appears.150 Importantly, identifying the victim is an
aspect which also appears in defixiones outside of Bath. At the site of Uley, for example,
defixiones also include the formulaic phrase ‘si vir si mulier si servus si liber’ ‘whether he is
man or woman, slave or free’.151 The two defixiones have notably similar phrasing with
respect to this aspect, highlighting the formulaic language. The formula ‘whether man or
woman’ predominates only amongst British examples of defixiones and rarely occurs outside
of Britain.152
As most curse tablets in Britain were mainly concerned with theft it would make sense for the
formula targeting unknown thieves to appear mainly on defixiones. In connection to Vulgar
Latin, the use of formulaic phrasing makes inscribing defixiones much easier, especially for
those who had not received a high level of education. Consequently, the correctness of
Bath’s defixiones when compared to Classical Latin may result from the use of formulaic
sentences. Hence Vulgar Latin terms can be placed within a sentence structure, resulting in
the ‘intruding’ aspect. When considering Bath’s defixiones, we should not immediately
regard them as completely original but as drawing instead on a common stock of language. In
150 Cunliffe (1988) 67. Cunliffe notes that variations of the ‘si vir si femina’ formula appear in around 11 examples of Bath’s legible Defixiones. 151 Britannia 10 (1979) p. 343, no. 3. 152 This is true of all the defixiones I have examined. The traditional Greco-Roman tablets tend to have a specific victim in mind when dedicating their curse. For example, one would know which rival in love they would want to bind. Conversely, identifying a thief is difficult as they are usually unseen when stealing. A catch-all formula is important as it identifies a variety of suspects to the deity ensuring the deity examines everyone.
50
this way, original names, stolen items, verbs and deities can be inserted into the formulae to
create defixiones. Often these inserted names and terms are spelt in a vulgar way or are
Celtic derivatives.
Notably, many words on Bath’s defixiones have vulgar spelling. For example, Tablet 65 has
inscribed on it: ‘Minerve de Suli Donavi’ (Tab. Sul. 65). On the tablet, the name Minerva
takes the dative singular case becoming Minervae as the tablet has been dedicated to
Minerva. However, because of vulgar spelling Minervae becomes Minerve; the same is true
for de which is a vulgar term for deae. Smith notes that it is common for the diphthong AE to
be pronounced or written with an E. Similarly, the ‘ER’ in certain words is replaced with an
AR.153 A large inscribed plate includes a list of names which highlight the use of AR rather
than ER (Tab. Sul. 30). The text lists both Patarnianus and Matarnus, terms which have
been identified by Cunliffe as vulgar forms of Paternianus and Maternus which are identified
by Smith as common cognomens in Roman Britain.154 The vulgarization of diphthongs such
as AE is not unique to Bath’s defixiones, as similar examples have been found all over
Britain.155
Furthermore, bilingualism has an important influence on the language which is inscribed on
objects such as defixiones. There is little evidence to suggest that Britons had an established
form of transliteration. Consequently, there is evidence of native British writing before
Roman occupation. So, Latin quickly became the primary form of writing in Roman Britain.
Nonetheless, ancient British culture and identity persisted through the influences that came
with Roman occupation. As is apparent even in Bath, Sulis retained her British name and
153 Smith (1983) 899-901. 154 Smith (1983) 898-899. 155 Herman (2000) 106.
51
existed in tandem with Minerva rather than having her identity completely usurped by the
Roman goddess. The most notable inclusion of British culture discovered on Bath’s tablets is
the use of personal names. British names such as ‘Veloriga’, ‘Brigomalla’ and ‘Aessicuna’
appear on defixiones and highlight the Celtic use of defixiones (Tab. Sul 53, 30, 98). The use
of Celtic proper names in inscriptions is usually indicative of native units within the Roman
army, or in Bath’s case, the cult site of a local deity.
David Mattingly notes that in pre-Roman Britain only small shrines were being built on
sacred spaces.156 With the arrival of the Romans, however, large temple complexes were
built, often on the location of traditional British cult sites.157 Importantly, the inclusion of
British names confirms Smith’s theory about British fluency in Latin. Attention should be
drawn to Tablet 14 as it has been considered to be ancient British written in Latin (Tab. Sul.
14). The fragmentary text combined with a lack of knowledge of ancient British makes
Tablet 14 untranslatable, but the tablet is an important source for bilingualism in Roman
Britain. It is notable moreover, that the text includes more than just a list of British names and
may reflect an attempt to write a curse completely in the British language.158 The
bilingualism present in Bath’s tablets shows that at least some of those depositing tablets into
Sulis’ spring were most likely Britons.
As noted above, one of the identifying features of the Latin used in Bath is the inclusion of
words which have been considered Vulgar Latin. These words have been identified ‘vulgar’
because they appear rarely, or never, in the works of Classical authors. The most common of
156 Mattingly (2006) 306. 157 The relationship between Roman religion and Celtic will be discussed later. 158 Cunliffe (1988) 128.
52
these is involare (‘to steal’); this verb appears in 30 instances of readable curse tablets.
Involare replaces verbs, such as corripere or auferre which are more commonly used to
describe the action of stealing.159 Similarly, the nouns employed on some of Bath’s tablets
reflect irregularities in the Latin used. For example caracalla is used to say ‘cloak’ rather
than a more common term such as lacerna. This is because terms such as caracalla have their
roots in the Celtic language which helps to explain their lack of prevalence in Latin.160 For
example, the caracalla was a hooded cloak which was most markedly worn by the Gauls and
Celts.161 Consequently, ‘caracalla’ fits into the list of Vulgar Latin terms and Celtic words
which are prevalent in Bath’s defixiones.162 Markedly, many of Bath’s tablets include a
variety of words which have been considered Celtic derivatives. For example, a small number
of Bath’s tablets use the word baro for man instead of vir which occurs more frequently in
defixiones. Only Tablets 44, 57 and 65 make use the word baro for man. Although the use of
this term is a rare occurrence, it does highlight the variety that bilingualism brings to Vulgar
Latin. Consequently, this usage shows that at least some of Bath’s tablets were petitions from
local Britons.
Defixiones from around the Ancient World.
Before discussing the judicial aspects of Bath’s defixiones, it is important to discuss what has
come to be considered as a ‘traditional’ curse tablet. David Jordan describes curse tablets as
‘inscribed pieces of lead, usually in the form of thin sheets, intended to bring supernatural
159 Oxford Latin Dictionary (1968) 212, 449. 160 Cunliffe (1988) 76. 161 Smith (1875) 240. 162 The caracalla is a Gallic hooded cloak which Cassius Dio claims originated in Germany and was made fashionable in Rome by the Emperor Aurelius Antonius Bassianus (aptly nick-named Caracalla due to his use of the cloak) during the second century CE. Bassianus made the caracalla standard issue for Roman soldiers (Cassius Dio. Roman History 79.3.3). The lacerna was shorter than the caracalla and hoodless. The lacerna had a variety of uses ranging from military wear to protecting Roman citizens from the rain (Pliny. NH 28.60.225).
53
power to bear against persons or animals.’163 Jordan provides a basic summation of what
curse tablets have been assumed to do. A victim’s name was inscribed onto the tablet by the
author of the tablet or defigens, and then deposited to begin the curse.164 However, this
chapter will demonstrate that defixiones were more diverse than Jordan suggests. Curse
tablets first appeared in Greece during the 5th century BCE and were very basic, usually
including only the name of the intended victim with limited references to verbs for binding or
cursing and only occasionally was a deity invoked.165 In many of these cases there was no
explicit verb to ‘bind’ or ‘curse’. A curse from Selinus emphasizes the simplicity of early
defixiones:
The tongue of Eucles and the tongue of Aristophanis and the tongue of Angeilis and the tongue of Alciphron and the tongue of Hagestratos. The tongues of the advocates of Eucles and Aristophanis. - SEG 26- 1113,166 CT. 49.167
The tablet is very basic, noting only the victim and the body part which was to be restrained.
The act of cursing is implied rather than specifically stated. By contrast, curse tablets found
during the Roman period boasted variations of intricate formulae (in both Greek and Latin)
with far more detailed curses than their predecessors did.168 A 4th century CE Greek curse
tablet, found in Rome, can be compared with the Selinian one above to emphasize the
development of curse tablets in the ancient world:
Eulamon, restrain, Ousiri Ousiri APHI OUSIRI MNE PHRI,… and archangels, in the name of the underworld one, so that , just as I entrust to you this impious and lawless and accursed Cardelus, whom his mother Fulgentia bore, so you may bring him to a bed of punishment with an evil death, and to die within five days. Quickly! Quickly!
163 Jordan (1985) 206. 164 Faraone (1991) 5. Faraone uses the term defigens to name the dedicator of defixiones. 165 Gager (1992) 5. 166 Supplementum Epigraphicum Graecum (1976) 26-1113. See this reference for the original Greek inscription. 167 Gager (1992) 139 168 Versnel (1991) 60. Versnel notes that isolating Greek from Latin defixiones is a hindrance as Greek material is heavily influential on Latin defixiones.
54
Spell: You, Phrygian goddess, nymph goddess. Eidonea [ADONAI] NEOI EKATOIKOUSE, I invoke you by your…, so that you may help me and restrain and hold in check Cardelus and bring him to a bed of punishment, to be punished with an evil death, to come to an evil condition, him whom his mother Fulgentia bore. And you holy Eulamon, and hold characters, and holy assistants, those on the right and those on the left, and holy Symphonia[?]. These things have been written on this (Eulamon, restrain, OUSIRI OUSIRI API OUSIRI MNE PHRI) tablet made from a cold-water pipe, so that, just as I entrust you to this impious and accursed and ill-fated Cardelus, whom his mother Fulgentia bore, bound, tied up and restrained, Cardelus whom his mother Fulgentia bore, that you may so restrain him and bring him to a bed of punishment, to be punished and to die an evil death, Cardelus whom his mother Fulgentia bore, within five days, because I invoke you by the power that renews itself under the earth, the one that restrains the circles and… - DT 155.169
The Sethian curse tablet is far more detailed than its predecessor and the tablet’s inscribed
contents are far longer than its forerunners, mostly due to the highly repetitive formulaic
language on the tablet.170 For example, it repeats several times ‘Cardelus whom his mother
Fulgentia bore’, a phrase which provided a means to identify the victim. Furthermore, later
defixiones often place the victim under the control of a divinity. In early examples, deities are
only rarely invoked on curse tablets and not in this much detail. By the 1st century BCE,
defixiones such as the Sethian one were well attested throughout the ancient world with
examples found in Egypt, Rome, Greece, Gaul and Britain.171 Examples of curse tablets have
been dated as late as the 4th century CE showing that they were used over a long period.172
Despite the development of defixiones, they all generally serve the same purpose. Namely,
that the outcome of a particular event is brought about in favour of the defigens. A desired
outcome was brought about by crippling or in rare cases, killing a rival. The use of
169 Audollent (1904) 208, no. 155. 170 This defixiones is noted as the ‘Sethian Tablet’ due to Wunsch’s argument that the Egyptian god Seth is represented on the tablet. On the Sethian tablet a figure is represented with a man’s body with a Donkey’s head which falls in line with the iconography associated with Seth. However, Gager also suggests that this could simply be a daimone tied to the horse racing. The Sethian tablet is important due to its good condition which makes it easy to read. The long inscription provides detailed insight into defixiones from the imperial age. 171 Ogden (2002) 10. 172 Ogden (2002) 210.
55
defixiones in the ancient world is varied as there were numerous reasons for a defigens to
dedicate a defixio.
As curse tablets developed they became far more specific and specialized. Daniel Ogden
notes that from the second century CE several new types of defixiones appear. These include
‘erotic-attaction’ curses, circus curses and athletic curses.173 Ogden notes that the
specialization of defixiones makes it easier to categorize them.174 For example, the Sethian
tablet fits into the category of highly specialized curses made to influence a specific, often
competitive, aspect of society. Importantly, Ogden goes on to note that defixiones such as
Bath’s begin to appear around the third and fourth centuries CE.175 As a result, the question
is raised as to whether Bath’s defixiones should be regarded exclusively as prayers for justice
or whether they fit into a larger context of the development of defixiones.
Within the ancient world, prayers and curses were not mutually exclusive aspects of society.
The tendency to separate the two is a modern construction of ancient religion. It is important
here to highlight the difficulty in differentiating curse from prayer as both have functions
which overlap. Both prayers and curses invoke the gods. Yet, as Watson suggests, ‘It is by no
means clear that a distinction was clearly recognised between praying and cursing.’176
Watson notes that prayers were typically addressed to particular deities, although curses were
not strictly bound to this convention.
173 Ogden (1999) 4. 174 Collins (2008) 65. 175 Ogden (1999) 5. 176 Watson (1991) 4.
56
The power of prayer in the Roman world.
The Romans’ relationship with their gods was contractual, legal and pragmatic. Ritual lay at
the centre of Roman religion and through sacrifice and prayer the pax deorum was
maintained.177 The gods would protect a person or state if their worship was properly
maintained.178 Prayer was an important aspect of keeping the pax deorum and accompanied
ritual practices such as sacrifice. John Scheid states that prayers were often made as
imperatives to be understood as official instructions by a magistrate.179 An example of prayer
can be found at the beginning of Marcus Varro’s Rerum Rusticarum. Varro states he will:
Et quoniam, ut aiunt, dei facientes adiuvant, prius invocabo eos, nec, ut Homerus et Ennius, Musas, sed duadecim deos consentes; neque tamen eos urbanos, quorum imagines ad forum auratae stant, sex mares et feminae totidem, sed illos XII deos, qui maxime agricolarum duces sunt.
And since, as they say, the gods aid those who ask, I will call on first, not the Muses, like Homer and Ennius, but those twelve councillor gods; but not those urban gods, whose gold gilt statues stand in the forum, six males and as many females, but those twelve who most of all are the guides of farmers. (Marcus Varro. Rerum Rusticarum 1.1-4).
Varro invokes the gods that benefit the growth of a farm in the countryside:
Primum, qui omnes fructos agri culturae caelo et terra continent, Iovem et Tellurem; itaque, quod ii parentes magni dicuntur, Iuppiter pater appellatur, Tellus terra mater. Secundo Solem et Lunam, quorum tempora observantur, cum quaedam seruntur et conduntur. Tertio Cererem et Liberum, quod horum fructus maxime necessari ad victum; ab his enim cibus et potio venit e fundo. Quarto Robigum ac Floram, quibus propitiis neque robigo frumenta atque arbores corrumpit, neque non tempestive florent. Itaque publice Robigo feriae Robigalia, Florae ludi Floralia institute. Item adveneror Minervam et Venerem, quarum unius procuration oliveti, alterius hortorum; quo nomine rustica Vinalia instituta. Nec non etiam precor Lympham ac Bonum Eventum, quoniam sine aqua omnis arida ac misera agri cultura, sine successu ac bono eventu frustration est, non cultura.
First, then, I invoke Jupiter and Tellus, who, by means of the sky and the earth, embrace all the fruits of agriculture; and hence, as we are told that they are the
177 Scheid (2003) 18. 178 Warrior (2006) 6. 179 Scheid (2003) 98.
57
universal parents, Jupiter is called "the Father," and Tellus is called "Mother Earth." And second, Sol and Luna, whose courses are watched in all matters of planting and harvesting. Third, Ceres and Liber, because their fruits are most necessary for life; for it is by their favour that food and drink come from the farm. Fourth, Robigus and Flora; for when they are propitious the rust will not harm the grain and the trees, and they will not fail to bloom in their season; wherefore, in honour of Robigus has been established the solemn feast of the Robigalia, and in honour of Flora the games called the Floralia. Likewise I beseech Minerva and Venus, of whom the one protects the oliveyard and the other the garden; and in her honour the rusticVinalia has been established. And I shall not fail to pray also to Lympha and Bonus Eventus, since without moisture all tilling of the ground is parched and barren, and without success and "good issue" it is not tillage but vexation. - Marcus Terentius Varro. Rerum Rusticarum I. 5-6.
Varro provides an example of how deities were invoked in Roman prayer, often for a specific
purpose. Although the deities are invoked together, they are called on for their individual
attributes which are beneficial in agricultural terms. For example, Minerva is called in her
role as the provider of olive groves. Tibullus states a prayer to protect his fields writing:
Di patrii, purgamus agros, purgamus agrestes: uos mala de nostris pellite limitibus.
Gods of our fathers, farms are cleansed, the farmhand cleansed. You gods drive evils from our boundaries! - Albius Tibullus. 2.1.17-20.
Specific gods are called to perform a protective function emphasizing the practical nature of
Roman prayers. Yet, prayers to the gods were also made for nefarious means.180 The historian
Velleius Paterculus notes that before his suicide the consul Merula prayed to the gods and
‘quos saepe pro salute rei publicae flamen dialis precatus erat deos, eos in execrationem
Cinnae partiumque eius tum precatus optime de re publica meritum spiritum reddidit.’
‘implored the gods to whom, as priest of Jupiter, he had formerly prayed for the safety of the
state, to visit their wrath on Cinna and his party. Thus did he yield up the life which had
served the state so well.’ (Velleius Paterculus. Compendium of Roman History Res Gestae
Divi Augusti 2.22. 2). Merula’s prayer demonstrates that there were darker aspects to prayer
180 Burriss (1930) 105.
58
as the gods could be invoked to avenge a perceived slight or to bring down political enemies.
In this respect, prayers are very similar to the defixiones which were found in Sulis’ sacred
spring. Prayers for revenge in the ancient world, although beneficial to the user, were harmful
to the victim, emphasizing a place where curses and prayers could intersect.
Notably, judicial defixiones such as Bath’s feature many of the characteristics of certain
gravestones. These stelai are inscribed with curses made to protect the graves from violation.
For example, one notes, ‘I Idameneus built this tomb to (my own) glory. May Zeus utterly
destroy anyone who disturbs it’ (IG XII, 1 737).181 Similarly, to judicial tablets such as
Bath’s, the gods are invoked only in response to an offence made to a dedicator, or in this
case, the deceased. The tomb violator is cursed only if he violates the tomb. In a judicial
sense, even if the violator is not caught by the local authorities, the deceased can be assured
that a deity will punish the violator’s transgression. Similarly at Bath, if the thief is not
caught, the goddess will surely punish him. Defixiones did have characteristics common to
prayers. A defigens clearly asks a deity to afflict a victim with various illnesses that assure
the suffering of the victim. Furthermore, the formulaic language of defixiones fits in with the
precise nature of Roman ritual and prayer. Greco-Roman and British examples of defixiones
often had specific means of invoking a deity on a tablet as well as the stock language used
when addressing the invoked deity. Importantly, the similarities between defixiones and
Roman prayer highlight the difficulty in distinguishing magic from religion as the two were
not mutually exclusive. In Rome, the use of defixiones acted as a private form of religion,
highlighting the variation of religious function in Rome.
Peter Salway notes that prayer for a specific objective to a specific god was central to ancient
181 Inscriptiones Gracae (1895) Volume XII, 1, 737. Also see Lattimore (1962) 109.
59
life.182 Such specificity in prayers was a common feature in the worship of local gods such as
Sulis. British defixiones highlight the relationship as Celtic deities were commonly revered
on a local basis. Regularly, in British defixiones a single Romano-Celtic deity was invoked to
retrieve a stolen object or, if the invocation fails, to punish the thief. In this sense, defixiones
appeared as a contractual relationship between the defigens and the deity who was invoked on
the tablet. In many cases the deity is promised a portion of the stolen goods, if they are
returned, as an ex-voto for the deity’s assistance (RIB 306).183 As previously mentioned, the
preoccupation with the restitution of thievery adds a legal characteristic to British defixiones.
Bath’s defixiones share the same prayer-like aspects as the greater corpus of British
defixiones.
Yet, defixiones did have unique aspects which are not apparent in most prayers. A large
proportion of defixiones address rivalry in the theatre or in sports. Often a defigens attempts
to bind the limbs of a race horse in order to stop the horse from racing.184 The act of binding
limbs and people is an aspect which is characteristic of defixiones. For example, a curse tablet
which was recovered from Rome asks a nymph, Eidonea, to restrain a victim of the curse:
I appeal to you Phrygian Goddess and Nymph Goddess Eidonea in this place that you restrain Artemios, also called Hospes, the son of Sapeda, and make him headless, footless and powerless with the horses of the Blue colours and overturn his reputation and victory. - Wunsch, Sethianische, no. 29.185
The anonymous dedicator hopes to influence the outcome of the races in his favour by
invoking Eidonea to restrain the limbs of the chariot driver Artemios. Similarly, the action of
182 Salway (1993) 484.183 This aspect is not even common amongst British defixiones, although the stolen objects may have been implicitly devoted to the deities upon their return. There is no evidence to suggest that altars were made as an ex-voto for fulfilled curses on tablets. 184 Gager (1992) 65. 185 Wünsch (1898) 40, no. 29.
60
binding is used to influence matters of love. Erotic defixiones were used to bind rivals of love
or to prevent the marriage of one’s love interest. One tablet from Greece notes:
I bind Theadora in the presence of one at Persephone’s side and in the presence of those who are unmarried. May she be unmarried and whenever she is about to chat with Kallias and with Charias-- whenever she is about to discuss deeds and words and business. Whatever he indeed says. I bind Theadora to remain unmarried to Charias and (I bind) Charias to forget Theadora, and I bind Charias to forget Theadora and sex with Theadora. - DTA 78.186
It is clear that the defigens wishes to keep the object of his affection, Theadora, away from a
suitor as well as trying to influence the suitor, Charias, into keeping his distance from
Theadora. Rather than simply paralyzing a victim’s body part, the act of binding in later
defixiones also binds the victim to the divinity tasked with cursing the victim. For example,
Theadora is bound to Persephone and Eulamon has had Cardelus entrusted to him. In these
examples, most of the curse’s power comes from a deity (or deities). Furthermore, the
dedication of people to deities was already an established convention of prayer. Livy
describes the dedication and sacrifice of Decius Mus along with the opposing army to the
Manes and Terra in order to secure victory for Rome (Livy. Ab Urbe Condita Libri 8. 6. 9).
The convention of dedicating a victim to the gods may have been an aspect which defixiones
later borrowed from other forms of religion such as prayer. As previously noted, early
defixiones simply identified a part of the body to be bound. However, defixiones differ tonally
as they were used to dedicate a personal enemy or rival to a deity often considered chthonic.
Furthermore, the defixiones differed from conventional religion and prayer as defixiones
would often be unleashed upon a fellow citizen. As such, the application and intention of
defixiones made them appear to be bad.
186 Gager (1992) 90, no. 22.
61
The formulae and subject of defixiones.
Defixiones whether British, Roman or otherwise, drew from a pre-existing stock of religious
formulae and verses.187 But these formulae were adapted to suit the purpose of the defigens
using a defixio. For example, imperial defixiones were used to address private concerns such
as rivals in love as well as the addition of binding victim’s limbs. Furthermore, the addition
of voces mysticae, foreign gods and other ‘magical’ phenomena to defixiones shows an
adaptation of traditional religion rather than an entirely new aspect of the supernatural world.
Additionally, the syncretistic nature of Roman religion meant that characteristics from
foreign cultures were added. Graf also appropriately notes that the verses and formulae found
in curses tablets can also be found in other religious activities such as prayer.188
Versnel has suggested that the terms ‘entrusting’, ‘dedicating’ or ‘handing over’ the victim is
a way of referring to the gods’ jurisdiction and influence.189 Additionally, Scheid notes that
the defixiones acted a form of ‘votive contract’ between the defigens and the deity invoked.190
The contractual aspect of later defixiones is similar to the legal nature of Roman religion. The
legal relationship is clear between Bath’s defigentes and the goddess Sulis. Hence, through
the dedication to a deity a curse received more supernatural clout due to the power which
came with the close proximity to a deity. By dedicating the victim to a deity, the defigens
gains assurance that the curse will work. Through both their similarities and differences to
conventional prayers, defixiones can be regarded as a specialized form of prayer.
187 Graf (1991) 197.188 Graf (1991) 197. 189 Versnel (1991) 90. 190 Scheid (2003) 105.
62
In contrast, Cunliffe recognises the lack of ‘religious language’ present in Bath’s
defixiones.191 Bath’s tablets are inscribed with a variety of stock formulae, but there is no
strictly set form on how to address Sulis through a defixio. It is important to note that
‘formulaic’ does not mean that each tablet was identical with no room for individuality;
rather, similar phrases were employed in the tablets’ creation. For example, some of Bath’s
tablets simply list names (Tab. Sul. 2), while others include intricate formulae designed to
cast a wide net over a variety of suspects (Tab. Sul. 44). The large variety of formulae
suggests a freer composition than the specific rituals or prayers which are common in Roman
religion. However, the function of ‘devoting’ a victim to Sulis is a ‘prayer-like’ aspect which
Bath’s curse tablets share with other defixiones and Roman Prayer.
Bath’s defixiones share many similarities with their cousins from other parts of the ancient
world. One tablet uses the term devoveo, meaning ‘I curse.’ (Tab. Sulis. 10). Similarly, tablet
99 uses the verb execro meaning ‘I curse’ (Tab. Sulis. 99); in both instances, the victim
curses the thief. This point is interesting as at least, to some petitioners, Bath’s tablets acted
in the form of defixiones. Yet, relatively few of Bath’s tablets curse the victim. Typically
punishment is placed in the hands of the goddess.
The majority of Bath’s defixiones are very similar in form to other defixiones. Bath’s tablets
were inscribed pewter sheets which were rolled or folded before been deposited into Sulis’
spring (fig. 13). Bath’s defixiones incorporated many features of curse tablets. For example,
several of Bath’s defixiones had curses inscribed with the Latin written in reverse (Tab. Sul.
191 Cunliffe (1988) 70.
63
44, 61, 62, 98, 99). The act of writing the curse backwards was common amongst curse
tablets all over the ancient world.192 By reversing the text, Bath’s defixiones become more
difficult to read achieving a form of ‘secret text’. Additionally, tablet 4 includes an example
of sympathetic magic. This refers to using the medium of the tablet or its deposition site to
influence the curse. So tablet 4 notes, ‘qu(i) mihi VILBIAM in(v)olavit sic liquat com(o)(do)
aqua.’ ‘May he who has stolen VILBIA from me become as liquid as water’ (Tab. Sul. 4).193
The ‘water’ refers to the water in Sulis’ spring; the defigens was using Sulis’ spring to
increase the power of the curse. The use of sympathetic magic, although not uniform, is very
common amongst Greco-Roman defixiones. For example:
As these names are chilled, so let the name, the soul, the passion, the mind, the spirit, the reasoning of Eros be chilled. - D.R. Jordan. ‘Agora’, no.1 (1975).
However, the most striking resemblance is the way in which the victims were punished on
Bath’s defixiones. Similarly to other defixiones, Bath’s defixiones will often transfer, or
dedicate, the thief into the goddess’ possession. In many of Bath’s defixiones, the defigens
chooses to ‘give’ the thief to Sulis rather than the item which was stolen. For example,
Tablet 61 notes, ‘Lovernisca d(onat) eum qui sive v(ir)i sive femina s(i)ve puer sive puella
qui mafortium involaverit’, ‘Lovernisca [gives] him who, whether man or woman, whether
boy or girl, who has stolen her cape.’ (Tab. Sul. 61). Similarly to the Sethian Tablet and the
erotic tablet, Lovernisca places an unknown thief into the power of a deity. However, the
thief was not bound as Cardalus and Theadora were but given to Sulis as if the person were
192 Ogden (1999) 29. 193 Tablet 4 is the only tablet in Bath’s cache which uses sympathetic magic in its composition. There is also doubt as to whether Vilbia refers to a woman or whether is a spelling error made by the author. There are no other curse tablets in Bath’s cache referring to people as the objects stolen. Cunliffe notes that the short list of suspects for a stolen love would also be shorter than the list on tablet 4. (see Cunliffe [1988] 4).
64
an item.194 The act of binding or giving a victim to a deity does not make him a divine slave
to the deity but rather, as previously mentioned, places him under the control of the deity.
Furthermore, if the thief is given to a deity such as Sulis, the deity’s attention is drawn to the
existence of the thief. In theory, the invoked deity would then pursue the thief and exact the
curse which was inscribed on the tablet.
Differences between Bath’s defixiones and examples from the
Mediterranean.
Conversely, Bath’s defixiones, along with other British variants, lacked several features
which were typical in imperial curse tablets. Primarily Bath’s defixiones lacked voces
mysticae, which were a series of distorted words which Gager notes can cover up to ninety
percent of a tablet.195 Fritz Graf has suggested that voces mysticae were a list of secret names
or epithets which belonged to the divinity invoked.196 The knowledge of these ‘secret names’
acted as credentials which added to power to the defigens and defixio. It has been argued that
the voces mysticae are used to bind a divinity to one’s will. Voces mysticae are present on the
Sethian defixio mentioned above. The tablet notes ‘Ousiri Ousiri APHI OUSIRI MNE PHRI’
with three of the words appearing in a mystic language (DT. 155).197 Similarly, defixiones
from Hadrumetum in North Africa highlights the use of voces mysticae stating,
‘HUESSE[M]IGAD[O]N IA[O BAUBO EEAEIE’ (DT 295).198 Notably, in both tablets the
voces mysticae appear in large writing. This aspect may function to add power to the voces
194 The verb donat ‘he/she gives’ is used on the majority Bath’s defixiones. Donat appears in its third person, singular present as Lovernisca refers to herself in the third person on Tablet 61. 195 Gager (1992) 7. 196 Graf (1991) 192. 197 Audollent (1904) 208, no. 155.198 Audollent (1904) 409, no. 295.
65
mysticae and seize the chosen deity’s attention. The voces mysticae were considered
particularly helpful when coercing a deity to curse someone.199
Many Greco-Roman defixiones included such voces mysticae. However, the lack of voces
mysticae in Bath’s tablets emphasizes the legal nature of the defixiones. Rather than binding
Sulis, she is being petitioned for help. The highly legal structure of Bath’s and British
defixiones generally make the use of voces mysticae and other religious language
unnecessary. In addition, the hyper-correct grammar found in British Vulgar Latin meant that
there was little room for strings of ‘secretive names’ or ‘unintelligible’ writing.
Consequently, British defixiones such as Bath’s have a distinct lack of voces mysticae found
on their tablets.
British defixiones differ tonally from Greco-Roman equivalents in that they lack the selfish
motives often present in traditional curse tablets. In British curse, tablets the identification of
the thief also highlighted the thief’s guilt in the eyes of the defigens. In a judicial sense,
British dedicators were identifying the people who had wronged them and placing them
under the deity’s judgement. Deities such as Sulis would in turn hunt down the criminal who
was brought to their attention. Consequently, defixiones such as Sulis’ highlight
characteristics of defixiones which can be interpreted as a judicial appeal to a deity for help,
displaying the versatile nature of defixiones and prayers and how the two categories may
intersect.
199 Graf (1991) 191.
66
Cunliffe notes that the ‘legalism’ in Bath’s defixiones is more striking than their
‘religiosity’.200 There was great variation of formulae in Bath’s tablets, from simply listing
the names of suspect thieves on the defixio to detailed inscribed instructions as to how Sulis
should punish thieves. Some tablets ask that the thief to pay with his blood while others do
not; such variety suggests that there was no strict ritual when approaching Sulis to recover
items. Furthermore, Cunliffe remarks on the legalism of Bath’s defixiones noting the use of
verbs which reflect a legal petition.201 Notably, the tablets include the use of legal terms such
as vindicere, to exact reparation or to avenge (Tab. Sul. 35) or satisfacere, to give satisfaction
or make amends (Tab. Sul. 94).202 In addition, catch-all phrases such as si vir si femina and si
servus si liber reflect a legal attempt to identify criminals to Sulis the judge. Notably, tablet
97 curses anyone who ‘keeps silent or knows anything about’ a stolen silver ring (Tab. Sul.
97). Those with knowledge of the ring’s theft were punished for not coming forward with
information. As previously stated, Bath’s defixiones are more notable for their legal than their
religious language. Consequently, Bath’s tablets acted as a specialized legal petition to a
goddess from those who have no legal representation elsewhere.
In many cases Bath’s tablets simply state ‘the name of the thief is given’ (Tab. Sul 15, 16).203
In Tablets such as 15 and 16, the stolen object and the name of the thief are not written down
but implied. Christopher Faraone suggests that simplistic defixiones such as Tablets 15 and
16 were accompanied by verbal incantation and during this incantation the name of the thief
and the stolen object were given.204 Faraone’s theory is possible at Bath as there are many
200 Cunliffe (1988) 70. 201 Cunliffe (1988) 70. 202 Satisfecere is commonly used in the formula sanguine suo satisfecere, to amend with one’s blood. This was a common formula across British defixiones. 203 Interestingly, the name of the thief is not present on the defixiones as is suggested. It has been noted by scholars such as Faraone and Gager that the name of accursed was invoked orally on the defixiones. This may have been the case with tablets 15 and 16. 204 Faraone (1991) 4.
67
examples of defixiones which simply include a list of names. For example, Tablets 30 and 51
simply list various names without a statement of cursing or a description of the lost item
(Tab. Sul. 30, 51).205 On such ambiguous defixiones it is difficult to discern whether or not
the thief was given to Sulis as there is no statement saying that the defigens was doing so. In
many cases, the defigens may not know the name of the thief or have a specific suspect in
mind. By stating ‘the name of the thief is given’ the defigens may be noting that although
they do not know the identity of the thief, Sulis may know the thief’s identity. Alternatively,
to Faraone’s hypothesis, defixiones such as 15 and 16 may simply present shorter means of
the thief is identified through long catch-all formulae such as those in tablets 32 and 36 (Tab.
Sul. 32, 36). When the names of potential criminals are listed the defigens likely thought it
was important that Sulis knew the identity of the criminal to be hunted. It was therefore not
always necessary to include elaborate inscriptions detailing the thief’s punishment.206 This
aspect of Bath’s defixiones is reminiscent of much earlier examples of curse tablets such as a
tablet found in Attica which notes, ‘All the choral directors and the assistant choral directors
with Theagenes, both the directors and the assistant choral directors’ (DTA 34).207 Similarly
to Bath’s tablet 30, a list of victims is given with the act of cursing implied. In this way, some
of Bath’s defixiones are reminiscent of earlier curse tablets with the defigens showing
knowledge of an earlier, more simplistic, tradition of cursing.
The reason thieves were placed under Sulis’ influence was because it provided the means for
Sulis to track them. For example Tablet 65 notes:
205 Cunliffe (1988) 146. Tablet 30 is inscribed on a pewter plate and is the only example of a plate used a curse tablet in Bath’s cache of defixiones. The medium of the plate is pewter keeping in line with Bath’s other defixiones. However, Cunliffe notes that the plate has the most elegant Latin from all the texts from the spring. 206 Faraone (1991) 4. Faraone suggests that both the oral and written aspects defixiones were developed at the same time. 207 Gager (1992) 49, no. 1.
68
Minerv(a)e de(ae) Suli donavi furem qui caracallam meam involavit si ser(v)us si liber si baro si mulier hoc donum non redemat nessi sanqu(i)n[e] suo.
I have given the thief who has stolen my hooded cloak, whether slave or free, whether man or woman. He is not to buy back this gift unless with his own blood. - Tab. Sul. 65.
In Tablets 61 and 65 the thief is committed to Sulis due to stolen items. Furthermore, Tablet
65 states that the thief will pay for his indiscretion. Rather than simply harming the victim,
Sulis was expected to act as a kind of supernatural law enforcer who would find the thief. If
the item cannot be returned to its owner, Sulis is expected to punish the thief, often by
making the thief reimburse the item with his blood. Consequently, the thief is placed under
Sulis’ jurisdiction and judgement as the thief is punished only if the stolen items are not
returned. Sulis’ invocation differs to that of other deities who are summoned specifically to
do harm to an unsuspecting victim.
Similarly to common defixiones, Sulis’ victims are subjected to various forms of gruesome
punishment. Furthermore, Bath’s defixiones subjected thieves to a large variation of curses
ranging from a simplistic list of names to detailed accounts of how Sulis should deal with the
thief and stolen goods. In this way, curses inscribed on Bath’s cache are suggestive of other
defixiones. Gager notes that commonly a defixio would seek to impact negatively the health
of the victim.208 Through a defixio, the victim was often subjected to various ailments such as
illnesses, impotency, sleeplessness and mental disability. Many of these afflictions are used
to punish Bath’s thieves. For example Tablet 54 notes:
N[on il[l]I p]ermittas nec sedere nec iacere [ne]c … a[m]bulare n[ec] somn[um nec] sanitatem [?cu]m quontocius consumas.
208 Gager (1992) 21.
69
You are not to permit [him] to sit or lie [or…. or] to walk [or] (to have) sleep [or] health. [since] you are to consume (him) as soon as possible. - Tab. Sul. 54.
The defigens asks Sulis to punish a thief with a range of ailments which impact the health of
the thief. Tablet 54 can be compared with an Athenian law tablet which notes ‘Let
Eutuchianos grow cold and not be in condition this coming Friday’ (Jordan, Agora no. 1).209
Notably, both tablets attempt to affect the health of victims highlighting similarities between
the two groups of defixiones.
Although Bath’s defixiones and ‘traditional’ defixiones share similar variations of
punishments, Bath’s tablets differ tonally. Bath’s defigentes ask that Sulis punishes her
victims gruesomely; in contrast with common Greco-Roman defixiones such punishment is
reactionary. In most cases the thief is only punished because he has stolen something from
the owner. In turn, a curse is incited on the behalf of the defigens. Versnel notes that illness
caused by supernatural forces was a powerful motivator to return stolen goods.210 In doing so,
the thief would have the chance to cure the affliction placed on him by an avenging deity.
However, if the item was not returned (the likely case) the defigens could take comfort in
Sulis’ punishment of the thief.
Furthermore, if the stolen goods could not be returned Sulis was commonly asked to
reimburse the item through the thief’s health. Amongst British defixiones it is common for a
stolen item to be paid for with the thief’s blood.211 Bath’s tablets are no exception to this
209 Jordan (1975) 246. The defixio is too fragmentary to know what the defigens wanted Eutuchianos to ‘not be in condition’. The curse’s location has led Jordan to believe that Eutuchianos was a rival in the law courts. 210 Versnel (1991) 84. 211 Versnel (1991) 84.
70
formula as many of Bath’s defigentes ask for blood as repayment for lost items. Likewise, In
Tablet 99 Deomiorix asks Sulis to punish the thief who robbed his house, asking, ‘inveniat
sanguine et vitae suae illud redemat’ ‘let him buy it back with (his) blood and his own life’
(Tab. Sul. 99). Similarly to the use of illness, the thief pays with his health to ensure the
defigens is compensated for his loss. In many cases the punishment is made to fit the crime.
Tablet 44 describes the theft of a bronze vessel and curses the thief to ‘spill his own blood
into the vessel itself’ (Tab. Sul. 44). Moreover, the act of filling the vessel with blood would
make the item less useful.212 Consequently, if the vessel remains missing it becomes tainted
and useless to other thieves who may acquire it.
Additionally, curses in ‘traditional’ defixiones were often final and irrevocable with no option
for the thief to repeal or remove the curse. In contrast, Sulis’ curses are often provisional. For
example, Tablet 45 notes:
Side A: Deae Suli.[… … ]. Is qu[i Side B: si servus si liber <si> qui<s> cumq[ue] erit non illi permittas nec oculos nec sanitatem nisi caecitatem orbitatemque quoad vixerit nisi haec ad fanum [
To the goddess Sulis… whether slave or free, <if> whoever he shall be, you are not to permit him eyes or health unless blindness and childlessness so long as he shall live, unless [he…] these to the temple. Tab. Sul. 45.
In this case the thief is only punished if the stolen item is not returned to Sulis’ temple.
Tonally, Bath’s defixiones lack the unprovoked attacks which characterize most defixiones. In
most cases, Bath’s defigens are more interested in the safe return of the stolen item rather
than making sure that the thief suffers. As Tablet 45 emphasizes, the thief’s health is affected
only if he does not return the stolen items to Sulis’ temple. In this way, Bath’s tablets can be
212 Tablet 44 is the only example of Bath’s cache of defixiones to include the violation of the stolen item.
71
seen to align with ‘judicial prayers’ rather than curses, as Sulis is petitioned to find a thief and
the item he or she stole, with punishment as the consequence only if the item cannot be
returned.
Furthermore, Cunliffe notes that Bath’s defixiones did include elements of sacrifice.213 As
previously mentioned, Bath’s thieves were dedicated to Sulis for punishment. This aspect of
the defixiones is reminiscent of the function of sacrifice as the thief’s life becomes Sulis’
property. The frequent mention of blood payment in Bath’s defixiones may also be a
reference to sacrifice. Both Sulis and the defigens benefit from blood sacrifice as the stolen
item is recompensed to the defigens while Sulis receives the blood of the thief, or the thief’s
livelihood. It is worth noting, however, that the idea of blood payment is not unique to Sulis’
defixiones. The action of ceding an item to the deity is a characteristic which is common
amongst ‘judicial prayer’ tablets.214 Within Britain it was common for a British defigens to
turn over their stolen property to a deity to retrieve. For example, a tablet addressed to
Nemesis states:
Domna nemesis do tibi palleum et galliculas qui tulit non redimat ni[si] vita sanguinei sui
Lady Nemesis! I give you this cloak and these shoes. May the person who has worn them not buy back (them) except with his own life and blood. -Magika, no. 100.215
Nemesis has been given possession over stolen property to exact an appropriate punishment
from the thief. Additionally, the tablet addressed to Nemesis tablet shares a common interest
with Sulis’ defixiones in ‘blood payment’ for a stolen item. The idea of buying an item back
213 Cunliffe (1988) 70. 214 Gager (1992) 175. 215 Versnel (1991) 198, no. 100.
72
with the thief’s blood is a common motif in British curse tablets with various deities asked to
reimburse an item with blood, for example, Mercury.216 In a legal sense, blood payment is an
appropriate punishment for an unreturnable item in a supernatural court overseen by a deity.
Bath’s defixiones conform to broader British characteristics as Bath’s defigentes will often
dedicate their stolen items to Sulis. Accordingly, Sulis is asked to retrieve a variety of items
from suspected thieves. For example Tablet 8 notes:
[d]eae Sulis donavi [arge-]ntiolos sex quos perd[idi] a nomin[i]bus infrascri[tis] deae exactura est Senicia(n)us et Saturninus <sed> et Ann[i]ola carta picta persc[ripta]
I have given to the goddess Sulis the six silver coins which I have lost. It is for the goddess to exact (them) from the names written below. Seniciacus and Saturninus and Anniola. - Tab. Sul. 8.
In Tablet 8, the sum of stolen money becomes Sulis’. As a result, the defigens has made the
slight against him also an offence to Sulis’ divinity as by dedicating the stolen objects to Sulis
they have become sacred objects. Consequently, recovering the stolen item was not only a
human issue but one of divine concern. As the thief has stolen sacred objects from a deity, the
crime is far worse than stealing from a human. Thus, the punishment is worse. The act of
stealing a deity’s property may justify the harsh punishments which are inscribed on the
defixiones. Ceding of stolen goods to a divinity also emphasizes the judicial aspects of Bath’s
defixiones. The stolen goods are clearly identified with Sulis, who was then expected to
recover the items from the thief. By identifying the item, Sulis knows what she is looking for.
Tablet 62 goes to great lengths to specify the stolen item stating, ‘perdedi la[enam]
[pa]lleum sagum paxsum’ ‘(my) Italian/Greek/Gallic cloak (and) tunic (Tab. Sul. 62).
Similarly to identifying victims, Tablet 62 ensures that all possibilities are taken into
216 Hassall, Wright (1973) 325, no. 3, Britannia 4 (1973) no. 3 p. 325.
73
consideration when identifying the stolen cloak. The item is also placed under Sulis’
jurisdiction through its cession into Sulis’ possession.
In many British examples however, the deity invoked on a defixio is often given only a
portion of the stolen item in exchange for help. For example, a tablet from Nottinghamshire
donates to Jupiter ‘one-tenth of the sum when he pays it’ (JRS 53).217 The act of giving a
portion of the stolen item is distinctive to British defixiones and can be seen in many
examples, (JRS 48218, no. 3, DT 106).219 Having said that, as previously noted by Cunliffe,
the spilling of blood or the dedication of the thief may have been payment enough.
Furthermore, many of Bath’s tablets compel the thief to return the stolen items to Sulis’
temple complex (Tab. Sul. 10, 32). The ex-votive offering may be implied when the stolen
item returns to the temple.
Similarly to the variety of formulaic language, the places where defixiones were deposited
varied. This depended on which forces were invoked on the tablets. Curse tablets were
deposited in graves, chthonic sanctuaries and bodies of water.220 These sites were notable for
their connection with the underworld and chthonic forces which were popular in Greco-
Roman defixiones.221 The same variation of deposition sites is clear all over the ancient
world. For example, in ancient Britain, defixiones addressed to various divinities have been
found in temples, wells and sacred springs. In many cases they were placed in the location
they were made to affect. For example a curse tablet made to bind chariot racers could be
217 Turner (1963) 123. 218 Audollent (1904) 159, no 106. 219 Wright and Richmond (1958) 150, no. 3. 220 Ogden (2002) 210. 221 Faraone (1991) 9-10.
74
placed on the hippodrome floor.222 There was no standardized placement of curse tablets;
they were placed where they were thought to work best.
Bath’s tablets are no exception to this as they were only found in Sulis’ sacred spring, which
lay within the centre of her temple complex. The placement of Bath’s defixiones makes sense
in legal terms, as placing the defixiones within the deities’ sacred boundaries helps to
legitimize the cession of potential thieves or, in Sulis’ case, stolen property over to the care of
the goddess. As Sulis’ sacred spring was the focal point of her power, placing the tablet here
meant that the defigens could maximise their connection to Sulis as well as the amount of
influence Sulis has on the curse. In addition, Ogden suggests that ‘prayers for justice’ such as
Bath’s could be placed in non-chthonic sanctuaries and be addressed to ‘ordinary
divinities.’223 However, as little is known about Sulis there is no easy way to discern whether
Sulis is a chthonic entity or not. In addition, it is clear that ‘prayers for justice’ did not
exclude chthonic entities as deities such as Hermes and Nemesis who have been invoked in a
similar way to Sulis (Magika, no. 100).224
In addition, Greco-Roman curse tablets differ tonally from British equivalents. In most
Greco-Roman tablets the victims are bound, impeding their success and therefore ensuring
(in theory) the dedicator’s own desired attainments. In this way defixiones had a selfish tone
as they existed predominantly to further the author’s own aims by inflicting pain on a victim.
Aside from personal gain, there is no other reason to dedicate a curse tablet. In most cases the
author has no previous grievance with the victim inscribed on the curse tablet. Furthermore,
222 Gager (1992) 18.223 Ogden (1999) 15. 224 Versnel (1991) 198, no. 100.
75
the defigens provides no statement defending the action of dedicating defixiones. The selfish
tone of curse tablets can further be emphasized in the interaction between the defigens and
deities who were invoked on the defixiones. Versnel notes that the divinities that were
typically invoked on defixiones were chthonic or had a strong connection to the underworld,
death or magic.225 John Gager supports Versnel’s consideration, noting that the most
commonly invoked deities on defixiones are Hermes, Hekate, Persephone, Hades and Gaia.
Latin equivalents included a very similar list of divinities noting Mercury, Manes, Jupiter,
Pluto, Nemesis and various water nymphs.226
As the previously mentioned defixiones demonstrate, curse tablets became far more syncretic
as time progressed, incorporating a wide variety of foreign deities. Egyptian and Jewish
deities were particularly popular in Rome.227 The predominant reason for including foreign
deities was that they embodied chthonic aspects that matched their Greco-Roman
counterparts. The partiality for chthonic entities parallels the dark nature of the deities and the
dark work of curses rather than a connection with death. Many of Britain’s defixiones are
syncretic in nature. Large portions of British defixiones are dedicated to Celtic deities, such as
Sulis, who are in turn equated to a Roman deity. Sulis’ connection with Minerva provides
one example.
The Relationship between defigentes and deities on curse tablets.
In Bath, however, the relationship between deity and defigens does not appear to be
traditional. In most cases the deities invoked on traditional Greco-Roman defixiones are
225 Versnel (1991) 64.226 Gager (1992) 12. 227 Collins (2008) 72.
76
coerced or ordered into carrying out the curses which were inscribed in defixiones.228 For
example a Greek defixio commands Hermes to bind a pimp and his working women:
I bind Theon, himself and his girls and his trade and his work-opportunities and his work and his speech and his deeds. Underworld Hermes, perform this act of restraint and keep reading this so long as these people live. - DT 52.229
Hermes is told rather than asked to perform the curse which was inscribed on the tablet. Most
notably the last sentence conveys to the reader that the curser expects Hermes to keep the
binding on Theon in place indefinitely in order to make Theon’s life miserable. Hermes is
coerced into cursing Theon as he cannot rest until he does the defigens’ bidding.
Consequently, we can see how a divinity can be forced into applying a curse to a victim. The
Sethian curse tablet similarly orders its comprehensive list of deities into binding the chariot
racer; notably the defigens asks the divinities to hurry with the cursing stating, “Quickly!
Quickly!” (DT 155).230 Graf notes that the coercing of gods has become a defining property
of magic.231 Although defixiones drew from common religious tradition, the coercion of
deities seems out of place when compared to the typical status quo between mortal and god.
The imperative tone of these defixiones shows little reverence for the divine status of the
deities. Furthermore, the deity is not promised a votive offering nor a token of thanks in the
event that the curse is not fulfilled. Consequently, defixiones highlight a strange reversal of
the religious ‘status quo’ with mortals making demands upon a deity. As such, the self-
serving tone of the defixiones is well represented through the defigens’ relationship with the
divinities he invokes.
228 Ogden (1999) 45. 229 Audollent (1904)87,no.52.Ogden(2002)215,no.177.230Audollent(1904)208,no.155.231 Graf (1997) 222.
77
In addition, the defigens’ demanding nature towards an invoked deity contrasts with the
highly pragmatic nature of Roman religion. Notably, Versnel states that defixiones use the
‘minimum criteria’ of prayer, a point emphasized through the imperative tone of common
curse tablets.232 As a result, the normally contractual nature between deity and mortal is
missing from traditional curse tablets. Deities such as Ousiris and Eidonia noted in the
Sethian curse tablet above are ordered to restrain a victim without any form of votum to the
deity. When considering Bath’s own defixiones Sulis is almost exclusively invoked on the
tablets; the only variation occurs when Sulis was invoked together with Minerva. Notably
there are few instances where specialized terms of imprecation are used in Bath’s defixiones.
Defixiones which use an imperative tone and language highlight a sense of urgency in the
curse on the Sethian tablet.233 The defigens wants Cardelus to be crippled within five days.
By compelling the deity rather than asking, the defigens gains assurance the curse will be
carried out faster. Faraone supports this view, noting that coercion of a deity was often used
as a prompt when the divinity did not act quickly enough.234 As previously mentioned, the
voces mysticae may have represented alternative names or epithets of deities. In turn, the
voces mysticae bring the defigens closer to the invoked deities.
Fritz Graf has suggested that a higher divinity must be invoked in order to manipulate lower
beings such as the daimones or the aoroi.235 Returning to the Sethian tablet, the tablet invokes
the god Osiris before noting ‘and archangels, in the name of the underworld one’ (DT 155).236
As the greater Egyptian deities Osiris, Mnevis and Apis are invoked, the defigens is also able
232 Versnel (1991) 61.233 See page Chapter 2, page 53 for a discussion on the Sethian curse tablet. 234 Faraone (1991) 194. 235 Graf (1997) 226. 236 Audollent (1904) 208, no. 155.
78
to invoke lower supernatural beings such as the archangels. Furthermore, as Osiris is the
Egyptian god who presides over the afterlife, the defigens is able to make a direct connection
to the ‘underworld one’.237 This hierarchy is apparent in other examples such as one which
invokes the ‘daimones’ in connection with Kronos and Typhon (DT 295).238 Importantly, this
supernatural hierarchy only became prominent during the Roman imperial age, as a result of
extensive fusion between Roman, Greek and other foreign religions.
Conversely, the coercion of deities and reversal of the status quo between god and mortals is
missing in British defixiones. In most cases the deities who are invoked on British tablets are
‘petitioned’, ‘addressed’ or ‘entreated’ rather than ordered. Furthermore, the idea of a
supernatural hierarchy is not a feature of British defixiones as only a single deity is
summoned. The beginnings of the texts on the tablets from Bath highlight their first contrast
with typical defixiones as the first words are important when establishing the relationship
between goddess and invoker. Versnel notes that supplication and submissive language are
rare in defixiones.239 Sulis’ dominance over the petitioner is made clear from the start of
many of Bath’s defixiones. She is invoked respectfully with full reverence given to her status
as a deity. For example, on Tablet 10 Sulis is invoked; ‘Docilianus Bruceri deae
sanctisiim(a)e’, ‘Docilianus (son) of Brucerus to the most holy goddess Sulis.’ (Tab. Sul. 10).
Tablet 10 highlights the relationship between Sulis and her petitioners. By addressing Sulis as
‘holy’, Docilianus clearly acknowledges Sulis’ divinity and in turn, her superiority to him.
Sulis is addressed with various titles which refer to her divine status. For example, Sulis is
invoked with titles such as ‘Goddess Sulis’ (Tab. Sul. 8), ‘Lady Goddess’ (98) and ‘Sacred
Majesty’ (35). In judicial tablets the deity retains the dominant role in the relationship with
237 Gager (1992) 70. Gager notes that the ‘underworld one’ is almost certainly a reference to the spirit of the deceased. 238 Audollent (1904) 409, no. 295.239 Versnel (1991) 61.
79
the defigens, a point which may be linked to the judicial role which is expected of deities in
judicial prayer tablets.
In other tablets the status quo is made more vivid. Tablet 35 notes:
Deae Sul[i] Minervae rogo [s]anctissimam maiestatem tuam u[t] vindices ab his [q]ui [fra-] [ude]m fecerunt ut ei[s per-] mittas nec s(o)mnum [nec [ ].[ ]. .[
To the goddess Sulis Minerva. I ask your most sacred majesty that you take vengeance on those who have done (me) wrong. - Tab. Sul. 35.
Although the petitioner is not named, clearly the individual is addressing Sulis. As such,
Bath’s tablets exemplify a traditional relationship between goddess and mortal. The author is
submissive to the goddess. By entreating the goddess to help, rather than commanding her, all
the power to punish the criminal lies with Sulis. As such, the defigens is entirely dependent
on the goddess when seeking justice. Through the defixiones Sulis presides as a judge either
by retrieving the stolen item or appropriately punishing the thief. As such, the defigens must
approach Sulis in respectful terms. The judicial aspect contrasts to the coercive relationship
between the deity and petitioner in traditional curse tablets. However, as with previous
defixiones, Bath’s tablets draw their power directly from Sulis.
Dedicators of Bath’s defixiones express a clear sentiment that they have been wronged by
someone. The injustices suffered by a victim are almost always mentioned on Bath’s tablets.
One tablet notes, ‘Minerv(a)e de(ae) Suli donavi furem qui caracallam meam involavit’, ‘To
Minerva the goddess Sulis I have given the thief who has stolen my hooded cloak’ (Tab. Sul.
65). The victim feels affronted through the theft of his cloak. The situation is important as it
provides a justification for invoking Sulis to avenge his wrong. This situation provides a clear
contrast to defixiones placed within other categories. Tablets such as Bath’s reflect a need to
80
avenge an offence rather than furthering one’s aims in politics or to cripple a rival. However,
Bath’s tablets should not be disregarded completely as defixiones, since distinguishing prayer
from curse and the occult from religion can be difficult.
Typically, Bath’s defixiones begin much like letters which the victims address to Sulis, often
stating their own names in the process. For example, Tablet 32 states ‘deae Suli Minerv(a)e
Solinus’, ‘Solinus to the goddess Sulis Minerva’ (Tab. Sul. 32). The author of a curse tablet
identifying himself is fairly irregular within the realm of defixiones. In doing so, the author
has revealed his name on the tablet. Such personal identification is contrary to the secretive
nature of defixiones. But in a prayer, especially one aimed at fixing a slight, it is important to
identify oneself as the supplicant in need of the deity’s aid. The defigentes lodge their
complaints with Sulis, but Sulis must carry out the punishment at her will and when she
chooses. The victim appeals to Sulis as one would to a magistrate, a point which highlights
the formulaic nature of Bath’s defixiones. Although such formulae keep in line with previous
curse tablets, they also highlight the aspects of Bath’s ‘prayers’ which make them different.
For example tablet 34 notes:
Deae Suli Minervae Docca dono numina tuo pecuniam quam [c.5 a] misi id est (donarios) (quinque).
Docca to the goddess Sulis Minerva. I give to your divinity the money which I have lost, that is five denarii. - Tab. Sul. 34.
Rather than an anonymous dedication, the victims often (although not always) name
themselves on the tablet, emphasizing their victimization at the hands of thieves as well as
absolving themselves of negative implications when dedicating defixiones. The defigentes’
self-identification on the defixiones is indicative of votive offerings as the dedicators of vota
were not usually anonymous. This highlights a difference between Bath’s defixiones and
81
those found in Rome because curse tablets were often anonymously deposited. In addition,
similarly to votive offerings it was appropriate for the defigens to name themselves when
dedicating a petition to the goddess. This further emphasizes the submissive tone of the
defigens’ approach to the superior deity.
Furthermore, Docca has identified the item which was stolen, in this case five denarii, so that
Sulis could attempt to find the money which was stolen. More commonly, however, the
stolen object was identified rather than the victim. Similarly to victims noted in defixiones
from Bath and other sites, the stolen objects are often ceded to the goddess. For example
Docca gives her five denarii to Sulis (Tab. Sul. 34). In contrast, regular defixiones were
dedicated anonymously with the name of the defigens purposefully omitted from the curse. In
this way the defigens avoids repercussions if the curse is discovered. By not naming oneself,
one could avoid counter curses and other legal ramifications from deities, another defigens or
the spirits of the deceased.
Additionally, a small cache of seven metal paterae were dedicated in Sulis’ spring and
addressed to the goddess. However, the honorifics used on the defixiones are different. The
tablets use flattering terms such ‘holy’ (Tab. Sul 10) and referring to her majesty (Tab. Sul
32). Terms such as these make the defixiones read much like a letter. Yet, the majority of
dedicated vessels use simpler titles when compared to defixiones such as, ‘Deae Suli’, ‘to the
goddess Sulis’ (Inv. no. 641)240 or ‘deae Suli Minervae’, ‘to the goddess Sulis Minerva’ (Inv.
no. 478).241 Furthermore, Sulis had many altars dedicated to her by visitors to her temple.
240 Cunliffe (1988) 56. 241 Cunliffe (1988) 57. The metal patera dedicated at Bath give no indication what they were dedicated for. They simply state ‘To the goddess Sulis’, consequently we cannot know if this is conjunction with the
82
Similarly to the paterae Sulis is addressed as a deity. The majority of altars were dedicated to
Sulis to ensure the health and safety of soldiers (RIB 143). These were often done by a
freedman who was grateful for his manumission from slavery and had ‘willingly and
deservedly fulfilled his vow’ (RIB 147).242 Vow fulfilment was one of the most common
reasons to dedicate an altar to Sulis, and this formula appears on the majority of Sulis’ altars.
Altars were also simply made as an offering to the goddess (RIB 151). Dedications such as
paterae and altars did not resemble personal letters to Sulis as the defixiones did.
Consequently, they lacked the legal and personal tone of Bath’s curse tablets.
Although dedications such as the paterae and altars were dedicated for different purposes
from Bath’s defixiones, the way in which they were dedicated to the goddess is very similar.
The goddess was approached respectfully and the item was dedicated with no act of coercion
or exploitation. Consequently, rather than simply being deposited into Sulis’ spring, Bath’s
defixiones may have served as dedications to the goddess. This ensured that Sulis received
the defigens’ petition for restitution and helped to pass the defixiones into the goddess’ realm
of influence.
Defixiones provided Bath’s residents and visitors with a medium to contact Sulis as a means
of recovering stolen objects and gaining restitution for that theft. The size of Bath’s cache of
curse tablets is much larger than those commonly found in other parts of the ancient world. In
most cases, curse tablets were deposited on an individual basis.243 The defigens would place
his defixio with little regard to places where other people may have deposited theirs. By
defixiones or for another reason. Aside from the defixiones and patera there were no other inscribed votums found in Sulis’ spring. 242 Cunliffe and Davenport (1985) 129.243 Ogden (1999) 15.
83
contrast, the deposition of curse tablets at Bath was focused on Sulis’ spring. Furthermore,
the huge number of defixiones found in Sulis’ spring suggests that the spring was a customary
and accepted place to petition Sulis.
The defigentes of Bath.
Through Bath’s defixiones we are able to discern the kinds of people who were primarily
visiting Sulis’ temple and dedicating tablets. Roger Tomlin notes that by making a
comparison between the names the tablets preserve and the items which are listed as stolen
we can discern who was visiting Bath’s temple complex.244
The items which were listed as stolen on Bath’s defixiones were small in size and easily
carried. There is no evidence to suggest that large amounts of property or sums of wealth
were stolen from Bath. These items were primarily personal belongings which had a practical
use. For example, seven of Bath’s legible tablets mention the theft of cloaks. Furthermore,
bathing tunics, bracelets, rings and gloves are also noted amongst the objects which have
been stolen. It is likely that garments would have been stolen by patrons who were visiting
Bath’s bathing complex which lay directly next to Sulis’ temple. Small amounts of coinage
are also mentioned amongst the items stolen from Bath; the amounts being 5 denarii (Tab.
Sul. 34) and various amounts of silver coins (Tab. Sul 54, 8, 98).245 The objects stolen from
Bath were easily portable, meaning that stealing them was not a hard task. From the various
items which have been noted as stolen we can discern the fact that many of the thefts at Bath
can best be described as ‘petty theft’. Mattingly describes the objects stolen as ‘relatively
244 Cunliffe (1988) 79. 245 Cunliffe (1988) 80.
84
minor losses’; it is true that the victims did not lose anything of great wealth.246
The humble nature of the objects stolen may tell much about the clients of Bath’s temple and
bathing complexes. The relatively small amounts of money that have been noted as stolen or
lost suggest that the victims were not particularly wealthy. Furthermore, the thefts could
have occurred due to a lack of proper protection of one’s personal belongings while bathing.
In the ancient world, if one was wealthy enough, the person could afford an attendant to
watch his possessions while bathing. Petronius mentions this aspect of Roman culture in his
Satyricon, when a slave is flogged for losing his master’s clothes at the baths (Petronius.
Satyricon. 30). Although it is not possible to know whether slave attendants were common at
Bath, it is reasonable to assume that the victims of theft as inscribed on the defixiones were
not able to afford an attendant to monitor their possessions. On the other hand, their loss may
simply have resulted from bad luck.
Additionally, there are instances when household objects were stolen from the property
where the victim was living. Tablet 99 notes ‘execro qui involaverit qui Deomiorix
hos(i)patio suo perdiderit.’ ‘I curse him who has stolen, who has robbed Deomiorix from his
house.’ (Tab. Sul. 99). We cannot know whether or not Deomiorix was a resident of Bath or
just visiting the bath complex, but it is likely that break-ins occurred in Roman Bath.
Similarly, Civilis insinuates that his ploughshare was stolen, ‘si (qui) vomerem Civilis
involavit ut an(imam) suam in templum deponat’ ‘If anyone has stolen Civilis’ ploughshare (I
ask) that he lay down his life in the temple’ (Tab. Sul. 31). Civilis’ tablet is not only notable
for the theft from his property, but it also tells us that Sulis had devotees outside of Bath. As
Civilis owned a ploughshare it is unlikely he lived within the town of Bath. Furthermore,
246 Mattingly (2006) 315.
85
Cunliffe has suggested that tablets 49 and 62 note the theft of horse blankets.247 The point
supports the evidence that people were traveling from the region outside Bath. Importantly,
the tablets highlight the fact that the rural population of Britain travelled to Bath to make
dedications to Sulis. One of the reason Civilis visited Bath may have been to gain restitution
for the theft of his ploughshare.
Conclusion.
Bath’s defixiones occupy an interesting position in Roman religion; they share many features
which are reminiscent of curse tablets from other parts of the ancient world. In both kinds of
tablets a defigens uses the tablet to affect a victim through supernatural means. This effect
was often achieved by appealing to a deity and placing the victim within the deity’s care and
influence. However, Bath’s tablets are more complicated than simply being used to gain a
supernatural advantage over a rival. In most of Bath’s defixiones the punishment inscribed on
the tablet is provisional; the victim is only punished if he has stolen a belonging or failed to
return that belonging. Tonally, Bath’s defigens make it clear that they are dedicating
defixiones because they feel they have been victimized through the theft of their possessions
and Sulis is invoked either to retrieve their items or to punish the person (or persons)
responsible for the theft. In this way, Sulis is invoked as a form of supernatural law
enforcement and judge, first finding the thieves and then punishing them. Consequently,
Bath’s defixiones could be considered as ‘judicial prayers’ or ‘prayers for justice’.
247 Cunliffe (1988) 80. Although a rare word, Cunliffe notes that caballarius, meaning horsemen, is derived from the Vulgar Latin word caballus which means horse. As caballarem appears as the object of theft it is unlikely that it translates to horsemen. Due to the mention of clothing earlier in the tablet, Cunliffe suggests a textile context for caballarem hence his translation into ‘horse blanket’.
86
Furthermore, the judicial aspects of Bath’s defixiones fit into a larger context of British curse
tablets as most British curse tablets address cases of theft similar to Bath’s.
Additionally, these tablets are an important source of Vulgar Latin in Britain. Bath’s
examples of Vulgar Latin are important as they provide evidence as to who Bath’s defigentes
were. The use of Vulgar Latin and the items listed as stolen highlight the lower socio-
economic background of Bath’s defigentes. In the under-policed fringes of the Roman world
less affluent individuals such as the majority of Bath’s defigentes may have felt they had
better representation by appealing to divinities such as Sulis rather than the local magistrates.
As such, Bath’s defixiones held an important legal role amongst Bath’s inhabitants, which
contrasts with the self-serving and vicious nature of defixiones which are found in other parts
of the ancient world. Bath’s defixiones were an important medium used for calling the
goddess Sulis to address injustices against her devotees.
However, distinguishing religion from magic is difficult in the ancient world. Although
Roman prayers could be used to ensure the growth of one’s crops, they could also be
employed to bring the wrath of the gods onto one’s enemies. The formulaic language of
defixiones is similar to the ritualized nature of Roman religion. In a way, defixiones were an
adapted form of private Roman religion and could be altered depending on context. For
example, the Sethian curse tablet addressed rival charioteers. British defixiones are
undeniably different from Greco-Roman equivalents but there are also many reminiscences
of earlier defixiones. Consequently, Versnel is right in assuming that the ‘prayers for justice’
such as Bath’s should be contained within their own category. However, this could be its own
category of specialized defixiones rather than an entirely separate aspect of Roman religion.
87
Chapter 3: The Cult of Sulis-Minerva and the goddess’ divine attributes. Although Sulis is regarded as the patron goddess of Bath, there is very little evidence
regarding her characteristics and how her cult functioned.248 Sulis has been described as a
healing goddess due to the hot springs around which her temple complex revolved. The
cache of 133 defixiones highlights a goddess who, by a defigens, was beseeched to exact
retribution for a perceived wrong which was often theft. Due to a lack of evidence, little else
is known about the goddess Sulis aside from her connection to Bath’s thermal springs and the
cache of defixiones. Consequently, hypotheses as to Sulis’ divine attributes are predominantly
drawn from physical evidence. Although originally a Celtic deity, Sulis became conflated
with the goddess Minerva through Roman interpretatio. Evidence from Bath’s temple
complex such as the Gorgon pediment confirms Sulis’ connection with Minerva. The
majority of evidence we have regarding Sulis comes from Roman religious iconography, as
much of the Sulis’ Celtic origins have been replaced by Roman religious culture. In this
chapter I will discuss Sulis’ role at Bath’s temple complex and how her cult may have
functioned. This chapter will argue that Sulis was a hybrid goddess, representing by British
and Roman religious ideals. However, due to the predominance of Roman religious
iconography and architecture and the lack of British material evidence, it can be difficult to
separate and identify British religious practice from the Roman. This chapter will also discuss
the British adaptation of Roman religious culture and iconography, such as the defixiones and
Gorgon pediment, and how this affects Sulis’ representation as a goddess.
The lack of evidence regarding cult activity at Bath.
Unfortunately, there is little to no evidence regarding the function of Sulis’ cult and the
248 Green (1996) 93.
88
characteristics of the goddess. However the large number of defixiones found in Sulis’ sacred
spring suggests that they were an important function of Sulis’ cult. A cache of defixiones on
Bath’s scale was highly unusual in the ancient world.249 As yet, there have been no other
locations where defixiones or dedications to the goddess Sulis have been found. Furthermore,
Sulis fits into a larger motif of Celtic deities being worshipped on a specifically local basis;
Sulis was revered only within the region of Bath as she was tied to the sacred spring of Bath.
There is not a lot of firm evidence for ritual activities taking place at Bath before the Roman-
British era.250 The study of Bath’s history has been impeded by the city’s continuous use.
Roman Bath was built over Pre-Roman Bath. Similarly, the modern city of Bath has impeded
the research of Roman town of Bath. Scholar’s such as Miranda Green and Barry Cunliffe
note that there is possible evidence of pre-Roman worship in the form of a causeway of
gravel approaching Sulis’ spring from the south-west.251 However, a single causeway does
not provide the most solid indication of Iron Age worship as it could simply indicate
settlement of the site. Nevertheless, a set of eighteen Celtic coins constitute the earliest
offerings deposited in Bath’s spring and can provide more solid evidence than the gravel
causeway. Yet, there has been some contention over the dating the coins. Cunliffe suggests
that it is reasonable to suggest that the coins can be dated to c.30 BCE.252 Lyn Sellwood dates
the coinage to c.51 BCE- c.30 BCE.253 Interestingly, the Celtic coin horde only accounts for
coin deposits a century before Bath’s Roman temple complex was built. 18 Celtic coins and a
gravel causeway do not provide solid evidence identifying pre-roman activity at Bath.
249 Cunliffe and Davenport (1985) 60, Gager (1992) 177, Mattingly (2006) 40, 311. The Temple of Mercury at Uley provides the only comparison, a similarly large cache of defixiones was found there. The majority of these have not been translated or published. Consequently, we do not have a precise number of how many have been found. 250 Cunliffe (1988) 177. Due to the construction of the Roman temple complex it is difficult to fix a precise date of Bath’s Iron Age origins. 251 Green (1995) 93.252 Cunliffe (1988) 279. 253 Sellwood (1980) 130.
89
Consequently, there is no solid record of pre-Roman Iron Age activity at Bath; this makes it
difficult to discern Sulis’ role in Britain’s pre-Roman world.254
Iron Age British offerings do not provide much in the way of clues as to how Sulis was
worshipped before Britain’s Roman occupation. Traditionally, Celtic religion before Roman
occupation was far less material as the ancient Britons did not have anthropomorphic
representations of their deities. The dwellings of Celtic deities were specific natural features
such as lakes, rivers and forests.255 Furthermore, the Britons had no form of writing leaving
no literary evidence of religious ritual or mythological cycles in Ancient Britain.256 Most of
our evidence for Iron Age British religion comes from Roman sources.
In his Bellum Gallicum, Caesar comments on the religion of the Celts. He discusses the rites
of the ancient druidic order noting that the druids were ‘divinis intersunt, sacrificial publica
ac private procurant, religions interpretantur’ ‘concerned with divine worship, the due
performance of sacrifices, public and private, and the interpretation of ritual questions.’
(Caesar. Bellum Gallicum 6.13). According to Caesar, the Celts observed strict ritual
observance similarly to the Romans. Gods and natural spirits were closely intertwined within
the private and public life of the Britons. A pragmatic approach was taken when the gods
must receive offerings in order to give divine blessings. Caesar elaborates on this transaction
when discussing human sacrifice: ‘quod, pro vita hominis nisi hominis vita reddatur, non
posse deorum immortalium numen placari arbitrantur’ ‘They believe, in effect, that, unless
254 This does not mean Sulis was not worshipped during Britain’s Iron Age rather; there is not enough evidence to know how Sulis was worshipped. 255 Webster (1986) 23. 256 Webster (1986) 23.
90
for a man’s life a man’s life be paid, the majesty of the immortal gods may not be appeased’
(6.16). Caesar goes on to discuss human sacrifice including the burning of victims in a wicker
man. By the time Sulis’ temple was constructed in the first century CE, however, the order of
the druids had been exterminated along with many of their practices such as human sacrifice.
Yet, many of the Celts’ traditional practices such as outdoor worship and pragmatic ties with
their gods were easily assimilated by the Romans.
Furthermore, it is important to note that the predominant contributor to our knowledge of
Celtic religion comes from the Roman military. Consequently, our view on ancient British
culture has a potential for distortion. Sources such as Caesar should not be taken at their face
value. W. Jeffrey Tatum notes that Caesar described his conquest of the Gauls, Germans and
Britons as a great expedition of the far northern parts of his known world.257 For example,
Caesar uses ethnography to emphasize his own Roman identity by contrasting it with the
foreignness of the Britons and Gauls. Caesar comments on their use of chariots, emphasizing
the backwardness of British warfare.258 When describing the Britons, Caesar creative an
image of ubiquitous tattooed warriors stating:
Omnes vero se Britanni vitro inficiunt, quod caeruleum efficit colorem, atque hoc horridiores sunt in pugna aspect; capilloque sunt promisso atque omni parte corporis rasa praetor caput et labrum superius.’
‘All the Britons, indeed, dye themselves with woad, which produces a blue colour, and makes their appearance in battle more terrible. They wear long hair, and shave every part of the body save the head and upper lip. (Caesar. Bellum Gallicum 5. 14).
Caesar describes the stereotypical British barbarian. His portrays a singular body of people
with the same cultural practices. In truth, the Britons were highly regionalized people each
257 Tatum (2011) 75. 258 Mattingly (2006) 34.
91
with their own cultural identity.259 Caesar’s distortion portrays the Britons like the rest of
Rome’s barbarian enemies with a large amount of embellishment. Like many Roman tales of
foreign countries and cultures, Caesar’s Gallic Wars was made for a Roman audience. While
praising his own campaigns Caesar also presents the Britons in a way which was
understandable to a Roman literary audience.260
The Romanization of ancient British religion.
Caesar used his Bellum Gallicum to paint the Gauls and Britons as worthy enemies of Rome
by emphasizing the differences between the Celtic peoples and the Roman. By creating a
worthy enemy in Caesar’s self-promotion of his conquests of Gaul and Britain. One of
Caesar’s purposes for describing the Celtic religion was to emphasize the difference of the
Celts when compared to Roman practices. Caesar particularly emphasizes this difference
through his description of human sacrifice (Caesar. Bellum Gallicum 6. 16). However, while
presenting the Gauls as alien, Caesar also makes them appear more civilized through his
description of their religion. Caesar notes that the druids were one of two groups of account
and dignity in Gaul (Caesar. Bellum Gallicum 6.13).261 Aldhouse-Green notes that Caesar’s
commentary is one of respect for a highly organised group of religious leaders.262
Caesar also comments on the Roman gods most frequently worshipped by the Gauls.263 He
notes that among the Gauls, Mercury was worshipped the most. However, Apollo, Mars,
259 Mattingly (2006) 51. 260 Mattingly (2006) 34.261 The other group Caesar describes are the knights. 262 Aldhouse-Green (2010) 33. 263 Although Caesar refers the Gauls when discussing the druids, he notes that the Gauls learn the rites of the Druids from the British and they would often to travel to Britain to learn the druidic practice (Caesar. Gallic
92
Jupiter and Minerva were also commonly worshipped with the same attributes as their
Roman equivalents (Caesar. Bellum Gallicum 6.17). The Gauls were exposed to Greco-
Roman culture before the Britons as Roman Britain lay on the fringes of the Roman Empire.
Consequently, Roman deities such as Mercury and Minerva had more time to take root in
Gallic culture. However, Caesar’s observations can be applied to the Britons. For example,
Mercury remained a popular deity within Roman Britain and took on various epithets and
epithets such as Mercury Andescocivoucos.264 Mercury’s predominance is also reflected
among defixiones as Mercury was the most commonly invoked god. Caesar’s description of a
Romano-Gallic Pantheon is problematic. As emphasized by Miranda Aldhouse-Green, the
Gauls were unlikely to have accepted the Roman Pantheon of gods so wholeheartedly as
early as Caesar’s conquest of Gaul.265 Caesar may have also been attempting to rationalize
the vague religion of the Celts through interpretatio. However, there must have been some
degree of religious hybridization between Celtic and Roman gods as by the first century CE
Roman ritual was the predominant form of worship in Britain. Although a god may have a
Celtic name, it was worshipped in a Roman style.266
Furthermore, Britons received a similar stigma even in the later centuries CE.267 Pliny
compares the Britons to the Persians noting: ‘Britannia hodieque eam adtonita celebrat tantis
caerimoniis ut dedisse Persis videri possit.’, ‘Even today Britain practices magic in awe, such
grand ritual that it might seem she gave it to the Persians’ (Pliny the Elder. Naturalis Historia
30.4). Britain’s ‘otherness’ is compared to that of the Persians as Pliny compares the Celtic
Wars 6.13). Consequently, we can assume that druidic rites would have been the similar all over the Celtic world. 264 Green (1986) 36. 265 Aldhouse-Green (2010) 34.266 Other forms of religion were not as prominent as Roman religion as they lacked the same amount of material evidence left behind by the Romans. 267 Aldhouse-Green (2010) 24.
93
druids to Persia’s Magi. Romans, such as Caesar and Pliny, clearly saw Britain as a foreign
country that was both edgy and dangerous.
Through interaction with the Mediterranean east and particularly the conquering of Roman
Britain, Celtic deities gained anthropomorphised forms through Roman interpretatio.268
Tacitus notes that interpretatio was ‘the interpretation of alien deities and of the rites
associated with them’ (Tacitus. Germania 43). Consequently, anthropomorphic
representations of Celtic deities were heavily influenced by Roman iconography. Miranda
Aldhouse-Green notes that interacting with the supernatural world in Britain became
solidified by the Romans.269 The Romans formalized religious architecture, ritual and the use
of repetitious religious iconography for religious identification. However Roman
interpretatio in Britain is complicated as it does not represent a complete Roman takeover of
British religion.
Although religious ritual became Roman, it did not mean that native British deities took
second place to the Roman pantheon.270 For example, the majority of Latin religious
inscriptions refer solely to British deities. David Mattingly provides figures which state that
169 out of 246 religious inscriptions refer singly to British deities.271 The Romans were
clearly aware of the power of local British divinities. Goddesses such as Coventina and
Sirona were the focus of extensive worship.272 Furthermore, many British gods were linked
with a Roman counterpart such as Sulis-Minerva. However, Mattingly emphasizes that the
268 Mattingly (2006) 218. 269 Aldhouse-Green (2008) 194. 270 Webster (1986) 41.271 Mattingly (2006) 215. 272 Green (1996) 99, 102.
94
name pairing of Roman and Celtic gods only made up 26 percent of Latin religious
inscriptions in Britain.273 Consequently, Roman deities could not completely replace the
deeply rooted British spirits.274
Sulis fell into a similar pattern as other Celtic deities; physical evidence suggests that by the
late first century CE Roman ritual was the predominant form of worship at Bath.275 By
keeping names such as Sulis, traditional culture was acknowledged but the deity was mostly
worshipped through a Roman lens. However, Bath’s defixiones and altars predominantly
invoke Sulis rather than ‘Sulis-Minerva’ or ‘Minerva-Sulis’.276 Tablet 65 is the only example
where the name Minerva is invoked before that of Sulis (Tab. Sul 65). Although Sulis was
conflated with the Roman goddess Minerva and received a Roman style temple complex, her
Celtic name was still invoked over that of her Roman one.
The role of the gods in Celtic mythology is also important. Similarly to Celtic religious ritual,
there is very little evidence which reveals the nature of the Celtic gods. The British deities
were less clearly defined than their Greco-Roman counterparts, as sculptural representation
and written dedications were alien to the ancient British.277 Consequently, it is difficult to
define the role of Celtic deities within a larger pantheon of gods such as in Greco-Roman
worship. Many Celtic gods seemed to share very similar attributes. For example, deities such
as Sulis and Coventina had their cults firmly rooted within a spring or well.278 Furthermore,
there was no clearly defined hierarchy to Celtic divinities, perhaps due to the highly
273 Mattingly (2006) 215. 274 Webster (1986) 42. 275 Cunliffe (1985) 65. 276 Cunliffe (1988) 70. 277 Salway (1993) 469.278 Green (1996) 99.
95
regionalized nature of Celtic cults as some deities were only worshipped in specific locations.
The point is significant, as when the Britons came to imitate the Romans’ form of worship,
classical representations of gods and inscriptions predominated. Sulis’ cult is influenced by
and expanded under a classical bias. Peter Salway notes that classical representation may be
used to express British themes which are foreign to the Roman world.279 When discussing an
ambiguous goddess such as Sulis, it is important to keep Salway’s theory under
consideration.
The structure and layout of Bath’s Romano-Celtic temple complex.
Classical representation was important in the initial formation of Sulis’ temple complex.
Sulis’ cult occupied a large temple area; the size was particularly unusual for the worship of
Celtic deities. Evidence such as dedicated altars and inscriptions from patrons suggests that
Bath’s temple had many benefactors.280 As a result, the temple may have been unusually
grand for a temple found in Roman Britain. Sulis’ tetrastyle temple was originally built
during the late Neronian or early Flavian principates (65- 75 CE) in the Classical style (fig.
14).281 The temple was originally a small Roman prostyle building which was raised on a
high podium and accessed via stairs. The cella was fronted by four Corinthian columns which
are also present on the large front pediment.282 The strictly Mediterranean form of Sulis’ first
temple complex suggests that ritual was altered to become largely Roman with little Celtic
influence.283 As Michael Lewis correctly notes the more religion was influenced by the
279 Salway (1981) 666. 280 Sulis’ various dedications are discussed later in this chapter. 281 On the architectural development of Bath’s temple see Cunliffe (1985) 33. For a chronological summary see page 177. 282 Cunliffe and Davenport (1985) 179. 283 Cunliffe and Davenport (1985) 179.
96
Romans, the more ambitious and elaborate the ornamentation would be on a temple.284
Romanization during the temple’s first period may explain how Sulis became connected with
Minerva as the Romans would have coupled a foreign deity with no physical definition to a
familiar Roman equivalent.
However, the development of Sulis’ temple suggests a change in the way Sulis-Minerva was
worshipped. From the second century CE Sulis’ temple underwent extensive alterations (fig.
15). The basic layout of Sulis’ temple was Roman in form. The temple favoured a Roman
orientation facing towards the east.285 The temple was built on a raised podium which was a
typically Roman characteristic.286 In addition, the temple’s prostyle structure was altered to
include a raised ambulatory which surrounded the temple (fig. 15-A). The temple featured a
large courtyard in front of the temple complex with a raised altar. The altar suggests the act of
outdoor sacrifice which fits with the Greco-Roman religious tradition. Furthermore, the
temple’s frontally located stairway was flanked by two smaller rooms which may have been
subsidiary shrines.287 The extra shrines may have been added to accommodate an increase of
offerings as Sulis’ cella did not increase in size.
The structure of Sulis’ temple indicates that ritual probably followed a characteristically
Roman format. A set of stairs led to the pronaos (porch) of Sulis’ temple where rituals would
have taken place.288 Behind the pronaos was the temple’s cella which would have housed
Sulis’ cult statue. Sacrifices would have been performed outdoors on the temple’s outdoor
284 Lewis (1966) 10. 285 Lewis (1966) 32. 286 Scheid (2003) 71. 287 Cunliffe and Davenport (1985) 180. 288 Scheid (2003) 71.
97
altar positioned in front of the temple complex (fig 15). Bath’s temple complex had a large
courtyard for crowds to gather for public sacrifices.
Traditionally, the Celts worshipped the forces of nature in the open air; they did not house
their gods.289 The Celts did not have temples until they made contact with the Greco-Roman
world.290 Through both Roman and Celtic influence British temples began to develop a
unique style appealing to both cultures. Although Bath is very ‘Roman’ in form, there is
some Celtic influence. Bath’s temple began to transform from a purely Roman temple to one
which included hybridized Romano-Celtic aspects. For example, the temple’s raised
concentric ambulatory created the characteristic ‘double square’ plan of the Romano-Celtic
temple (fig 15-A).291 The ambulatory surrounded the temple’s cella creating an outdoor
characteristic which matched Celtic sensibilities.292 Celtic deities were traditionally
embodiments of natural features, for example, Sulis and her spring.293 It was not uncommon
for Romano-Celtic temples to incorporate these natural features as central features in the
layout of their temple complex.294 As a result, the temple’s orientation around a sacred spring
was not uncharacteristic for a Romano-Celtic temple.295
A spring which housed a goddess was commonly placed as the central focus of the goddess’
temple complex; this is clear in Sulis’ case as not only the temple, but the city of Bath was
orientated around her sacred spring. Other goddesses, such as Coventina, were housed in
289 Webster (1986) 107 290 Lewis (1966) 4. 291 Lewis (1966) 8. 292 Mattingly (2006) 282. 293 Lewis (1966) 4. 294 Mattingly (2006) 307. 295 Allason-Jones (1996) 107.
98
temples with a similar orientation.296 In Coventina’s case, her well replaced the cella of the
temple.297 In both Sulis and Coventina’s cases, their respective bodies of water acted as the
primary receptacles for their offerings.
Although much of Sulis’ temple remains hidden beneath the modern township of Bath,
excavations have yielded some evidence of Bath’s ritual functions. For example, a large
outdoor altar occupied the centre of the temple complex’s courtyard (fig. 15). The altar was
highly decorated with images of deities such as Bacchus (fig. 16-A), Hercules (fig. 17-C) and
Jupiter (fig. 17-A). Interestingly, the altar had no images of Minerva or Sulis. The altar was
most likely dedicated to Sulis as there is little evidence to suggest alternative deities were
venerated at Bath.298 In addition, the altars’ placement in the centre of courtyard meant that it
would have been used in outdoor sacrifices. Sulis’ altar conforms to the highly decorated
forms of altars found throughout Roman religion. Within Bath there are various examples of
inscribed altars dedicated to the goddess.299 Furthermore, outdoor sacrifices keep in line with
the conventions of Roman religion as sacrifices were made in open space in front of the
public.300 In addition, Bath’s main altar was located in front of Sulis’ temple complex. The
altar increased in size from the second century CE onwards emphasizing the temple’s
296 For Coventina see Allason-Jones (1996) 107. For the goddess Sequana see Green (1996) 91. And lastly, Sirona is dicussed in Green (1996) 102. These three goddesses provide alternative examples to Sulis when discussing temple complexes revolving around springs or wells. 297 Allason-Jones (1996) 107.298 There are three curse tablets which have been dedicated to deities other than Sulis (Tab. Sul 33, 53, 97). They tablets ask for Mars and Mercury to carry out the curse. However, there is not enough evidence to state confidently that Mars and Mercury were fully worshipped at Bath, although as Bath has not been fully excavated there still may be a chance more evidence could be uncovered. There are also representations of other deities such as Luna (fig. 21) and Diana, Cunliffe (1985) 131, present at Bath. 299 I discuss these in chapter 2 page 82- 83. 300 Scheid (2007) 263.
99
growing popularity.301 An increase in offerings and altar size would reflect the increase in
popularity of Sulis’ temple complex.
The Celtic features of the temple complex were less noticeable than the Roman. Sulis’
prominence at Bath is the clearest display of the temple’s British heritage. Furthermore, as
previously stated, Sulis’ spring remained the centre of worship. Similarly to the Ancient
Britons, offerings continued to be deposited into Sulis’ spring. Although whole human
figures were rare in Celtic art, the head was a common motif in Celtic art and religion.302
Bath’s Gorgon conforms to the Celts’ artistic representations of heads (fig. 18). The Gorgon
has oval shaped eyes as well as a flat nose. The figure has long hair and a beard which
completely surround the Gorgon’s head (fig. 18). The Gorgon strongly resembles head
representations of Celtic gods. Celtic representations of the head such as Bath’s Gorgon are
common when portraying deities.303 For example, small representations of males show a
masculine head with a beard and flowing long hair.304
Bath’s Gorgon pediment: A hybridization of Roman and Celtic religion.
Bath’s temple did retain some traditional Celtic iconography. The Gorgon pediment also
reveals a distinctly Celtic identity (fig. 18). Primarily the Gorgon is male not female, as is
emphasized by the Gorgon’s moustache, beard and furrowed eyes brows.305 Bath’s Gorgon is
consistent with depictions in ancient Greek art. The Gorgon is bearded with long hair and a
301 Cunliffe and Davenport (1985) 183. 302 Green (1996) 138. 303 Green (1986) 46. 304 Henig (1995) 39. 305 Richmond, Toynbee (1955) 104.
100
pair of wings similar to Gorgons depicted in Greek art.306 However, the Sulis Gorgon lacks
the monstrous and feminine appearance of other Gorgons. Bath’s Gorgon depicts the face of
a solemn elderly male. The representation of Bath’s Gorgon has led to alternative ideas about
what the pediment portrays. One hypothesis is that the Gorgon represents Bath’s legendary
King Bladud in the guise of a divinity.307 Traditionally, Bath was founded by Bladud who
made the city’s hot springs suitable for bathing. When discussing the founding of Bath,
Geoffrey of Monmouth notes that Bladud placed Bath under the protection of Minerva
(Geoffrey of Monmouth. Historia Regum Britanniae 2.30). Additionally, Bladud was
responsible for the eternal fire at Bath which Solinus also mentions in his brief description of
Bath’s layout (Solinus. Collectanea rerum memorabilium 22.10).308 The Gorgon’s wings are
consistent with Bladud’s reputed ability to fly.309 However, wings are not common attributes
in representations of Celtic deities.310 Care should be taken regarding the source’s reliability.
For example, Bath’s temple complex can only be dated as far back as Rome’s Neronian age
(c. 65 CE).311 Consequently, it is more likely the Romans were responsible for Minerva’s
presence at Bath rather than Bladud. Archaeology would suggest that the pre-Roman bath
was solely a sacred spring with no large buildings.
Other interpretations of Bath’s Gorgon include the head of Oceanus.312 The ‘wave-like’ hair
of the Gorgon is a representation of the ocean’s hair. The tritons emphasize this oceanic
aspect. However, I disagree with John Hind’s interpretation, as Bath’s Gorgon is far more
suited to iconography associated with the Gorgon. For example, there are snakes tangled in
306 Marinatos (2000) 48. 307 Fear (1992) 223. 308 It should be noted that Geoffrey wrote his history of Britain’s kings during the 1130s. 309 Interestingly, Bladud met a death similar to Icarus as he crashed into Apollo’s temple at Trinovantum and was instantly killed (Geoffrey. 2.30). 310 Fear (1992) 223. 311 Cunliffe and Davenport (1985) 178. 312 Hind (1996) 358.
101
the Gorgon’s beard and the head is flanked by wings. Oceanus is usually portrayed in Celtic
and Roman art with horns and beard sprouting seaweed rather than wings.313 Furthermore,
the presence of a Gorgon is more fitting with Sulis’ connection with Minerva. Firstly, the
pediment on which the Gorgon resides is ornamented with iconography associated with
Minerva (fig. 19). In Roman representations of Minerva, the goddess is typically portrayed
with her Gorgon-decorated cuirass.314 When Minerva was introduced in Britain she arrived
with the imagery of a goddess wearing her helmet, Gorgon-cuirass and wielding a spear.315
Several British goddesses such a Brigantia adopted Minerva’s iconography when their
images were created.316 However, unlike Sulis, goddesses such as Brigantia have become
regarded as British goddesses of war. It has been contentiously suggested that the Gorgon is a
male representation of Sulis.317 Yet, Sulis’ connection with the female goddess Minerva, as
well as the goddess’s address as ‘dea’ makes this unlikely.
The temple also had a focus on solar imagery which appealed to both Celtic and Roman
religious sensibilities. The sun had an important role in both Roman and Celtic religion as a
giver of life.318 Bath has solar imagery throughout its temple precinct. For example, Bath’s
Gorgon pediment is crowned with a small representation of the sun (fig. 18).319 Cunliffe
suggests that this may mean that the Gorgon may in fact be a representation of the sun god
Sol.320 Green also notes that the Gorgon’s hair is wavy evoking images of the sun’s rays.321
However, aside from the pediment’s crown there is no more evidence to suggest that the
313 Brilliant (1974) 139. 314 Shearer (1996) 81. 315 Green (1996) 36. 316 Green (1996) 36. 317 Cunliffe and Davenport (1985) 116. 318 Webster (1986) 32. 319 The connection between the sun and healing at Bath’s temple complex will be discussed later in this chapter. 320 Cunliffe and Davenport (1985)116. 321 Green (1996) 97.
102
Gorgon is a representation of the sun. Due to the large amount of iconography associated
with Minerva, it is far more likely the pediment boasts a Gorgon of Celtic origin.
Furthermore, the Façade of the Four Seasons portrays Luna and anthropomorphic
representations of the seasons which were closely tied to the phases of the moon (fig. 20,
21).322 The Façade is known for its anthropomorphic representations of the four seasons.
The solemn male-like appearance of Bath’s Gorgon contrasts with Gorgons in Greco-Roman
myth. Yet, the Gorgon has snakes tangled in his beard keeping in line with the iconography
which is associated with gorgons. Like many of Bath’s features, the Gorgon pediment is a
fusion of Celtic and Greco-Roman artistic features. The subject matter, the Gorgon, is a
figure from Roman myth. However, the Gorgon is rendered in a Celtic style. The Gorgon’s
lentoid shaped eyes, wedge-shaped nose and frowning brow portray a clear Celtic style in the
portrayal of the Gorgon.323 Moreover, the pediment’s focus on the Gorgon emphasizes the
Celtic religious focus on the head. The ancient Britons held the human head in high reverence
as they believed the head held the essence of someone’s being.324 Furthermore, the head was
also believed to house a person’s soul, making it an important symbol of divine power and
closely linked to the Celtic spiritual world. Consequently, Bath’s Gorgon occupies a
monumental position in Bath’s temple architecture. In addition, the Gorgon is encircled by
two floral wreaths. The use of vegetal motifs is popular amongst Celtic and Roman
symbolism which displays status and divinity.325 However, the Gorgon appears to be the only
artistic feature on the pediment, and at the temple complex, which is a strictly Celtic artistic
feature.
322 For Cunliffe’s discussion on the Façade of the Four Seasons see Cunliffe (1985) 123- 128.323 Cunliffe and Davenport (1985) 115. 324 Webster (1986) 39.325 Green (1996) 122.
103
Bath’s Gorgon does have Classical themes sculpted in. The serpents in the beard and the
wings are an acknowledgement of Classical themes and help to identify this Celtic head as a
Gorgon.326 The Gorgon is consistent with imagery associated with Minerva and her myths.
The circular wreaths around the Gorgon are representative of the Gorgon’s place on
Minerva’s aegis or shield (fig. 19). Although linked with Minerva through myth, snakes were
also potent symbols of healing within Roman religion.327 Consequently, the Gorgon’s snakes
would not have been out of place at a healing sanctuary such as Bath. However, there is no
serpent iconography anywhere in the temple complex aside from those in the Gorgon’s beard.
The presence of snakes in the Gorgon’s beard may not have provided a connection Minerva
Medica. Instead, the snakes may simply be iconography that is typically associated with a
Gorgon. Bath’s Gorgon would have created a powerful image to visitors entering Bath. The
large pediment is a symbol of Bath’s prosperity as well as an acknowledgement of Minerva’s
patronage at Bath. Importantly, artistic representations such as Bath’s Gorgon reflect a
primarily Romanized cult which incorporates surviving Celtic motifs and nods to Celtic
religious sensibilities. The pediment was most likely made to aggrandize Bath while its
iconography notifies visitors that the temple complex they were entering belonged to Sulis-
Minerva.
Sulis’ Minervan attributes: The conflation of two goddesses.
Bath reveals a goddess who has both Roman and Celtic identity. As previously mentioned,
Sulis gained her anthropomorphised form from the Romans hence equating her with the
326 Cunliffe and Davenport (1985) 115. 327 Ogden (2013) 317.
104
goddess Minerva. The only representation of Minerva found at Bath is a large bronze head
which once belonged to a larger statue (fig. 22).328 The hole in the back of the head marks the
place where Minerva’s Roman helmet once sat (fig. 23). The presence of Bath’s Gorgon
pediment is consistent with Minerva’s Mediterranean iconography (fig. 18, 19).329 Both the
Greek Athena and the Roman Minerva were equipped with an aegis bearing the Gorgon’s
head. Furthermore, the Gorgon pediment includes various iconographic features which
highlight a connection with the Roman Minerva (fig. 19). The pediment boasts a helmet on
the bottom left, matching Minerva’s established iconography and her role as a goddess of
strategy (fig. 19). Furthermore, the Gorgon’s outer wreath boasted a small owl (fig. 19).
Similarly to the military helmet, the owl complements Minerva’s traditional iconography.330
A Roman temple was usually built to commemorate a single attribute of a god, as is the case
with Minerva Medica’s temple built on Rome’s Esquiline Hill to venerate the goddess’s
patronage of medicinal craft.331 However, Bath gives no indication that Sulis was worshipped
for just one specific attribute.332 Aside from her connection with Minerva, there is no record
of Sulis having any epithets connected to her name. Conversely, Sulis fits into a larger group
of Celtic deities revered on a local basis.333 Sulis’ power as a deity was linked with her sacred
spring. Accordingly, there has been no evidence of Sulis’ cult outside of Bath. Localized
deities such as Sulis tend to have their power centred in one region.334
328 Although the bronze statue may have once been the cult statue of Sulis-Minerva, this is as of yet, unproven. 329 Shearer (1996) 7. 330 Cunliffe and Davenport (1985) 115. 331 Scheid (2007) 154. 332 Roman deities were commonly worshipped for a range of reasons rather than one characteristic alone. 333 Henig (2004) 224. 334 Salway 1981) 666.
105
It is also necessary to examine Minerva’s aspects and to compare these to Sulis. There is very
little evidence linking Minerva and Sulis aside from inscriptions. The most prominent
evidence we have are the defixiones which often refer to the goddess as ‘Sulis’ or ‘Sulis-
Minerva’(Tab. Sul 10, 32, 34, 35). Similarly to the defixiones, many of the votive paterae are
dedicated to ‘Sulis-Minerva’.335 In addition, Bath’s bronze head of Minerva as well as the
Gorgon pediment provide physical evidence regarding Sulis’ relationship to Minerva.
However, this is the only definitive evidence we have connecting the two goddesses.
As previously mentioned, Bath has distinct iconography in relation to Minerva; for example,
the presence of the Gorgon and the Victories on Bath’s pediment. In this context, Ann
Shearer discusses the Gorgon’s protective properties. In many cases the Gorgon appears on
the outside of buildings.336 In this way the Gorgon’s gaze acts as a protective ward against
evil spirits.337 Shearer’s hypothesis may have some weight as Sulis’ judicial attributes fit well
with a Gorgon who has protective properties, as Bath was a place of healing. The defixiones
portray a goddess who was believed to have protected her devotees by avenging the wrongs
which were inflicted on them. Furthermore, many of Sulis’ altars were dedicated as votive
offerings for the safety of an individual. For example, one such altar states ‘(D)eae Suli pro
salute et incolumita(te) Mar(ci) Aufid(i) (M)aximi c(enturionis) leg(ionis) VI Vic(tricis)
(A)ufidius Eutuches leb(ertus) v(otum) s(olvit) l(ibens) M(erito)’, ‘To the goddess Sulis for
the health and safety of Marcus Aufidius Maximus, centurion of the Legio VI Victrix,
Aufidius Eutuches, his freedman, willingly and deservedly fulfilled his vow’ (RIB 143).
Various other altars repeat the same formulae; a freedman dedicates an altar to entreat Sulis
335 Cunliffe (1988) 55.336 Shearer (1996) 68. 337 Wilk (2000) 150.
106
to safeguard his former master or to thank Sulis for her protection (RIB 144, 147). Similarly
to asking Sulis to punish a victim, one could ask for the goddess’ protection to keep someone
from harm.
Within Rome, Minerva was an important patron of arts and crafts. Furthermore, the creation
of defixiones was tied with industry and crafts.338 Minerva’s patronage of crafts is relevant at
Bath as the goddess could be seen as a patron of medicinal craft. Likewise, medicinal craft
was widely accepted within Celtic godhood; deities such as the Irish Brigit emphasized
healing and healing craft.339 In addition, Miranda Green notes that the martial traits of a deity
such as Minerva were transmuted into guardianship against disease.340 At Bath this is visible
in the various martial motifs such as the helmets and winged Victory figures on the Gorgon
pediment. The Gorgon is heraldically flanked on either side by two winged Victory figures
standing on globes; this is consistent with the Greek Athena’s relationship with Nike (fig 18,
19).341 It is clear that the sculptors used a large amount of Roman imagery when portraying
Sulis. Consequently, it may be likely that Sulis came to incorporate many of Minerva’s
characteristics and functions. Cunliffe discusses the potential for the winged Victory figures
as symbols of victory over illness.342 In conjunction with Sulis’ thermal springs, Cunliffe’s
hypothesis may be correct. Sulis’ association with Minerva Medica may support Cunliffe’s
theory as Minerva was a healing deity in the Roman mind.343 The symbolism would have
been important when promoting the sanctuary as one of healing.
338 Shearer (1996) 78. 339 Webster (1986) 110. 340 Green (1996) 94. 341 Shearer (1996) 8. 342 Cunliffe and Davenport (1985) 115. 343 Cunliffe and Davenport (1985) 115.
107
Iconography particular to Minerva such as the helmets and winged Victories feed into
Minerva’s martial characteristics. Sulis’ martial characteristics can be applied to her
punishment of victims inscribed upon the defixiones as she is often asked to deliver brutal
punishments to the victims described on the tablets. Bath’s defixiones emphasize the goddess’
martial abilities in her punishment of the victims inscribed on the curse tablets.
Sulis’ role as a healer is emphasized by most scholars due to the medicinal properties
attributed to Bath’s hot springs.344 As well as a sacred spring, Bath had a large bathing
complex attached to the southern end of the temple. Sulis, the Celtic embodiment of Bath’s
hot springs was equated with the goddess Minerva. Sulis’ clear popularity promoted her
interpretation as a primary deity such as Minerva. Other than Sulis, in most cases Celtic water
divinities were usually interpreted by the Romans as nymphs. For example, Coventina
appears in the form of a nymph in Carrawburgh.345 Although Sulis was not identified as a
nymph, Bath was not without aquatic imagery. For example, two figures believed to be
tritons are believed to be located on either side of Bath’s Gorgon pediment (fig. 24).346
However, due to the poor condition of the pediment it is difficult to see the tritons.
Consequently, it is difficult to confirm that there are representations of tritons on the
pediment. Although they would be fitting with Bath’s spring and bathing complex, they don’t
match the Minervan iconography of the rest of the pediment.347 Bath’s sacred springs as well
as its bathing complex were significant features of the temple. Consequently an important
connection between water and healing may be drawn.
344 Cunliffe (1988) 102, Green (1995) 93, Shearer (1996) 84. 345 Mattingly (2006) 215. 346 Richmond, Toynbee (1955) 99. 347 Smith (1983) 122.
108
There is, however, a lack of ex-voto offerings at Bath. It was common for a supplicant to
dedicate ex-voto offerings to the divinity once they had been cured. However, the dedication
of votive offerings may have fallen out of prominence by the time. Normally, these ex-votive
offerings took the form of the body part which was cured.348 At Bath, however, only a pair of
ivory breasts and a bronze breast could be considered ex-votive offerings. Furthermore, the
altars dedicated to Sulis do not mention any of the dedicators being cured by Sulis’ divinity.
Importantly, Cunliffe and Green both note that anatomical ex-votive offerings fell out of
popularity in Rome from the end of Rome’s Republican period.349 Consequently, the lack of
ex-votive offerings should not be taken as a suggestion that Bath was not a healing shrine.
A possible connection to the sun: Sulis’ solar attributes.
As well as Bath’s large number of ex-votive offerings, there was some iconographic evidence
suggesting that Bath had a healing function. For example, Sulis may have been associated
with the sun: the top of Sulis’ Gorgon pediment is crowned with a sun, and Green argues that
Sulis’ name was philologically linked to the proto-Indo-European word for sun.350 The proto-
Indo-European word for sun is sehaul.351 Within greater Celtic religions the sun held
important healing properties.352 These are illustrated by the cults which worship the god
Apollo.353 The presence of the sun would help to explain the goddess’ heated pools and
emphasize their divine healing properties. Religiously, the sun was also important to the
Celts. The spoked wheel was a symbol which was commonly associated with the sun in
348 Scheid (2003) 100. 349 Cunliffe (1988) 360, Green (1995) 95.350 Green (1996) 33. 351 Mallory, Adams (2006) 128. 352 Green (1991) 20. 353 Green (1991)
109
ancient Britain.354 A pewter vessel dedicated to Sulis also had an image of the sun engraved
onto the bottom, but other than this vessel, Sulis appears to have received no other votive
offerings with sun motifs.355 Unfortunately, physical evidence does provide much support to
the theory Sulis was a solar goddess.
The sun and water also share important links. Green notes that the water, sun and healing
were all closely tied each other in the Celtic world.356 Green’s hypothesis does carry some
weight. Visually, water reflects the light of the sun making it sparkle in the light.
Furthermore, both the sun and the water are essential to life. To the people of the ancient
world, the sun’s influence was present in hot springs such as Bath’s.357 Due to the sun’s heat,
the connection to Bath’s hot springs would make sense. Both the sun and the hot spring
renew life with their heat. Arguably, the hot springs’ main source of healing was their natural
minerals.
The presence of the sun symbolism would normally contrast with the dark nature of
defixiones. In this case, though, the cult of Sol Invictus had a prominent judicial role in
Roman religion from the third century CE, derived from his role as an all-seeing god.358 J.
Strubbe notes that, similarly, sky gods such as Helios, Zeus Olympios and Selene were
invoked in curses which protected tombs.359 These gods were known for their judicial
qualities and their ability to see all crimes. If one disturbed the remains in the tomb then
354 Green (1996) 148. 355 However, as Bath is still not fully excavated it is likely that there are more offerings to be found. Aside from the defixiones, Sulis’ spring has not yielded a large sum of offerings. Consequently, it is hard to draw evidence from votive offerings alone. 356 Green (1991) 20. 357 Green (1991b) 119.358 Stephenson (2009) 36. 359 Strubbe (1991) 46.
110
similarly to victims on Bath’s defixiones, the tomb desecrator would be punished. In addition,
Helios and Sol are often invoked in curse tablets which are categorized as ‘prayers for
justice’.360 Such judicial qualities attributed to the Greco-Roman sun religion are consistent
with those demonstrated by Sulis. Sulis oversaw the retribution of crimes against her
followers as Sol did with his.
Furthermore, another of Bath’s pediments includes an image of the goddess Luna (fig. 21).
Cunliffe suggests that pediment originally stood above the Façade of the Four Seasons.361
Luna can be identified by her iconography. The goddess has copious amounts of hair and is
draped. The moon crescent behind Luna’s head and the whip used to drive her ox-drawn
chariot (fig. 21). Luna’s presence at Sulis’ temple is plausible, and highlights the judicial and
healing characteristics attached to Sulis as a goddess. Luna was an important deity within the
Roman Pantheon; she provided a means to keep track of the year via the phases of the moon
(Virgil. Georgics 1. 5-6). Importantly, Luna was also a potent symbol of fertility, growth and
life (Cicero. De Natura Deorum 3. 119). Luna’s fertility symbols are similar to the
rejuvenating attributes found in both the sun and the water.
Bath’s solar and lunar iconography fits into a greater architectural trend occurring during the
third and fourth centuries CE. Depictions of the moon were never far away from those of the
sun. Sol and Luna often appeared opposite each other on Roman architectural features. The
Arch of Constantine which was built c.315 CE is an appropriate comparison to Bath’s own
360 Strubbe (1991) 46. 361 Cunliffe and Davenport (1985) 126. Cunliffe only guesses at the pediment’s original location. The Luna pediment was originally found in three blocks. When matched with the Façade the proportions fit. The moon also had a strong connection with the changing of seasons. Consequently, Luna’s presence above the personification of the four seasons would be appropriate.
111
celestial iconography.362 Sol and Luna appear on the arch as tondos with Luna on the arch’s
western side and Sol on the arch’s east. Cunliffe suggests a similar layout within Bath’s
temple precinct, noting that the Façade of the Four Seasons faced a southern portico with an
image of Sol on it.363 Similarly to the Arch of Constantine, Bath’s portrayal of Sol and Luna
depict the two deities in tandem. Sol Invictus became a mainstream deity and cult in the third
century CE.364 The proposed dates for the prominence of Sol Invictus’ cult match the
development of Bath’s temple complex.
Although Sulis may have similarities with Luna such as healing properties as well as judicial
aspects defixiones, there is no solid evidence to suggest that the two were closely connected.
As Green and Cunliffe have emphasized, there may have been a Celtic connection through
the sun through the heat of the spring but at this point it is difficult to draw to a solid
conclusion about Bath having aspects of a sun cult. Bath’s representations of the sun are
minimal and images of Luna and the seasons were probably made to beautify the temple
complex.
Sulis as a judicial deity.
Sulis’ role as a magistrate is further emphasized in her altars. Many of Sulis’ altars are
dedicated on the completion of a vow. Various vows were made by freedmen asking Sulis to
protect their previous masters (RIB 143, 144, 147). However, most altars do not describe
what kind of vow has been completed, only that it was completed. For example, one such
362 Stephenson (2009) 156. 363 Cunliffe and Davenport (1985) 127. 364 Stephenson (2009) 76.
112
altar states ‘Priscus Touti f(ilius) lapidaries(s) cives Car(nu)tenus Su(li) deae v(otum) s(oluit)
l(ibens) m(erito)’, ‘Priscus, son of Toutius, stonemason, a tribesman of the Carnutes, to the
goddess Sulis willingly and deservedly fulfilled his vow’ (RIB 144). On the completion of a
vow, an altar is dedicated to Sulis in order to mark this. In addition, a defixio from Sulis’
spring is dedicated as a sanction against perjury. The tablet notes:
Uricalus Do[c]ilosa ux[or] sua Docilis filius suus Docilina Decentius frater suus Alogiosa nomina<a> eorum qui iuraverunt <qui iuraverunt> ad fontem deae Suli(s) prid(i)e idus Apriles quicumque illic periuraverit deae Suli facias Illum sanquine sui illud satisfacere.
Uricalus, Docilosa his wife, Docilis his son and Docilina, Decentinus his brother, Alogiosa: the names of those who have sworn at the spring of the goddess Sulis on the 12th of April. Whoever has perjured himself there you are to make him pay for it to the goddess Sulis in his own blood. - Tab. Sul. 94.365
Tablet 94 emphasizes Sulis’ role as a judicial deity. Those who have broken oaths are
punished in a fashion similar to thieves. Dedications are made to Sulis when oaths are
completed while perjurers were punished.
Sulis’ healing attributes can be compared with her role as a destroyer which is emphasized by
Bath’s defixiones. Sulis clearly shows aspects of duality as a part of her cult. Sulis’ hot spring
would have had important healing properties for those who bathed there. However, the
defixiones display Sulis’ more harmful side. While the goddess was able to restore health, she
was more than capable of taking health away. Due to the inscriptions on Bath’s defixiones,
we know that Sulis’ worshippers expected her to cripple the health of and even take peoples’
lives (Tab. Sul 10).366 In contrast, there is no written evidence at Bath which notes healing as
a part of the temple’s function; this is implicit due to Bath’s hot springs.
365 Tablet 94 is currently the only tablet in Bath’s cache which punishes perjury. 366 In Tablet 10 Sulis is asked to take the life of the thief Brucerus. Furthermore, the goddess is also asked to inflict Brucerus with both impotency and insomnia. Tablet 10 is a strong reminder that while Sulis could heal, she was also capable to of crippling health or taking life.
113
Sulis’ discretion when smiting was dependant on people’s behaviour. For example, criminal
activity such as stealing or perjury was punished by Sulis. Yet, Sulis was kind to those who
were sick. However, Sulis is not the only deity who incorporates aspects of duality in their
characteristics. It is common for many deities to both help and harm mortals.
Most deities have a range of attributes which help to define their characters. In many cases
gods are able to both give and take. Gods who had seemingly beneficial attributes were also
capable of wrath. For example, the god Apollo was capable of both healing and
destruction.367 As a god with solar attributes, Apollo presents an appropriate comparison to
Sulis. Contrasting with this, Sulis’ cult seems to acknowledge a variety of different
characteristics.
Religious evidence from Bath’s sacred spring.
The development of Sulis’ sacred spring highlights the continuing change of the goddess’
religious rites. During Bath’s second period (c. 2nd century CE), Sulis’ sacred spring was
enclosed within a vaulted chamber (fig. 25). Previously, the spring was visible from the
raised temple as well as from the sacrificial altar giving an open atmosphere to the temple
complex with the spring being easily approachable by the public (fig. 14). The spring’s new
chamber restricted water and human access to the spring. However, the reason behind the
vault’s construction is unclear. Green notes that the addition of the vault may have been an
attempt to make Sulis’ cult more remote and mysterious, therefore increasing the religious
367 Ferguson (1989) 15.
114
experience.368 The transition to a vaulted chamber would have also placed more emphasis on
the spring as a place intended for worship. The entrance to the spring was directly connected
to Sulis’ altar. Accordingly, any religious ceremonies in the courtyard would have direct
access to the spring. As previously mentioned, Sulis was traditionally the spirit of the spring
and the vaulted chamber would have provided a far more appropriate ‘home’ for the goddess
than an open spring. Consequently, the vaulted entrance mediated a transition between the
secular world and sacred world. Within the vaulted chamber it was possible to contact the
goddess as is seen on Bath’s defixiones.
Interestingly, the vaulted spring’s construction date coincides with Cunliffe’s proposed date
for when the defixiones began to be deposited at Bath. This change in Sulis’ religion may
reflect the chthonic aspect which is associated with defixiones. As discussed in previous
chapters, the deities typically associated with defixiones have chthonic attributes.369 However,
the presence of defixiones at Bath should not be taken as a definitive answer to Sulis’
attributes. As previously mentioned in this chapter, sources of water such as rivers and
springs were important to Celtic religious practice. Similarly important was the role the earth
played in conjunction with these springs. Places where the water flowed from the earth were
held in special reverence.370 The Celts believed that these were the homes of powerful spirits
such as Sulis. Sulis has a clear connection between the earth and her aquatic attribute; her
connection to the chthonic realm is highlighted by her spring.
It is likely that Bath’s vaulted spring was built to commemorate a goddess whose popularity
368 Green (1995) 94. 369 See Chapter 2 page 74 of this thesis for a larger discussion on the connection between chthonic deities and defixiones. 370 Webster (1986) 107.
115
has risen over the years. Furthermore, the presence of a vaulted chamber draws attention to
the centre of Sulis’ power and could be seen as a part of the temple’s beautification rather
than an acknowledgement of Sulis’ chthonic nature. It is far more likely that Sulis’ spring
was enclosed as a way to add more grandeur and mystique to Sulis’ temple complex. It is
unlikely defixiones were deposited in Sulis’ spring because she was a chthonic goddess. It is
more likely that the defigentes were approaching Sulis as the most powerful divinity in Bath’s
region. By invoking Bath’s patron goddess on the defixiones, the defixiones were more likely
to gain more supernatural clout.
The spring’s development highlights a more personal form of worship towards Sulis. The
discovery of a life size bronze head portraying Minerva suggests that there were sculptural
representations of Sulis around her precinct.371 It is unknown where the bronze statue of
Minerva once stood or whether it once stood in the temple’s cella.372 While it is the finest
surviving sculpture from the temple, it is impossible to say whether it was the cult statue.373
In fact, there may not have been a cult statue: in many cases, Romano-Celtic temples took a
secondary place to sacred springs and groves, and small offerings to the deity were typically
deposited in the spring rather than near a cult statue.374 The inclusion of four square piers
scattered around the edge of the spring may indicate the inclusion of cult statuary within the
sacred spring (fig 25-B, C, D, E). The four square bases and three circular bases within the
spring may have supported statuary alternating between columns.375 The square bases were
371 There is no evidence to suggest that Sulis was represented and anthropomorphized before Roman occupation of Britain. Consequently, Sulis’ iconographical features likely matched those of Minerva. Only the head remains with the rest of the statue (including Minerva’s helmet) missing. 372 Cunliffe and Davenport (1985) 114. 373 There is no evidence to suggest whether the head once belonged to Sulis’ primary cult statue or another representation of the goddess from another part of the precinct. 374 Lewis (1966) 5. 375 Cunliffe and Davenport (1985) 42.
116
made by placing rectangular blocks on top of each other and sealing them with mortar (fig
26).376 Cunliffe notes that the bases had well tooled joints to fit the blocks together.377
However, the tops of the bases do not have marks or groves to hold a statue in place. Cunliffe
suggests that this is because the top layers of the bases have been removed.378 Statuary in
Sulis’ spring may have been dedications made by Sulis’ worshippers to honour the goddess;
their placement in the sacred spring would not be out of place considering the spring was
Sulis’ primary receptacle for offerings.
Sulis’ spring was likely the centre point of the goddess’ religious complex. In British
tradition Sulis was the embodiment of the spring, as was reflected in the temple orientation.
The temple complex, courtyard and bathing complex were built around the spring.
Furthermore, votive offerings made to the goddess were exclusively deposited within the
spring. Consequently, the spring was the focal point for communicating with the goddess and
it would make sense for an angry defigens to lodge complaints with the goddess here.
However, combined with the restricted access to the spring, the inclusion of cult statuary
would have increased the majesty of the spring as Sulis’ place of worship. Consequently, the
religious aspects of Sulis’ shrine were changed drastically. The chamber was accessible only
through a single entrance from the temple’s courtyard. Cunliffe has noted that this door may
have been utilized exclusively by priests to approach the spring while the public approached
from the south viewing through three windows (fig 25-A).379 These restrictions on Sulis’
spring would have made the space more sacred, as only temple staff was allowed access to
the spring. Green suggests that religious officials restricted access to the temple to exercise
376 Cunliffe and Davenport (1985) 42. 377 Cunliffe and Davenport (1985) 42. 378 Cunliffe and Davenport (1985) 42.379 Cunliffe and Davenport (1985) 50.
117
more control over the goddess.380 Votive offerings from the public may have been made
through the three windows at the southern end of the spring. Alternatively, the priests may
have placed all offerings within the spring. However, Cunliffe’s consideration is guesswork
as there is no evidence to suggest that the public were not allowed to enter the spring through
the entranceway, although enclosing the spring did mean that space was restricted, limiting
the number of people who could visit.381 The windows may have been placed so the springs
could be seen from the bathing complex which lay adjacent to Sulis’ temple.
The sacred spring’s entranceway connected to the courtyard of the temple (fig. 15). The
entranceway may have connected the spring to the altar so that religious processions could
reach the spring from the altar. Considering the spring’s prominence at Bath and Sulis’
strong ties to the spring, it would make sense to include the spring in religious rituals.
The only written evidence on Sulis comes from inscriptions dating to Bath’s period of Roman
occupation.382 These inscriptions predominantly appear on votive offerings found within the
spring such as the paterae and the defixiones. Less frequent are the inscriptions found on
tombstones and altars. Many of Bath’s inscriptions highlight the fact that Bath had many
benefactors who dedicated altars to Sulis. Many of these altars were gifted to the goddess and
do not specify whether they were dedicated to commemorate the fulfilment of a vow or as a
monument to the deceased. For example, an altar simply notes, ‘Quintus Pompeius Anicetus
Suli’, ‘Quintus Pompeius Anicetus to Sulis’ (RIB 148). It is not entirely certain why Quintus
has dedicated his altar to Sulis. It was not uncommon to dedicate altars to deities without
stating a reason. For example, an altar in Chedworth simply notes ‘To Lenus Mars’ (RIB
380 Aldhouse-Green (2004) 200. 381 Cunliffe and Davenport (1985) 50. 382 Cunliffe and Davenport (1985) 178.
118
126). These altars were probably dedicated by a deity’s more affluent worshippers.
Furthermore, inscriptions found on tombstones and altars are particularly useful as they
highlight some of the ritual functions at Sulis’ temple. For example, one tombstone notes
‘Dis) M(anibus) G(aius) Calpurnius (R)eceptus sacer dos deae Sulis vix(it) an(nos) LXXV
Calpurnia Trifo sa l(i)berta(a) coniunx f(aciendum) c(uravit)’, ‘To the spirits of the departed;
Gaius Calpurnius Receptus, priest of Sulis, lived 75 years’ (RIB 155). Furthermore, there is
evidence of an augur at Bath on a dedicated altar noting, ‘(D)eae Suli pro salute et incolumite
(te) Mar(ci) Aufid(i) (M)aximi leg(ionis) VI Vic(tricus) (A)ufidius Eutuches Leb(ertus)
v(otum) s(olvit) l(ibens) m(erito).’, ‘To the Goddess Sulis, Lucius Marcius Memor, augur,
gave this gift’ (RIB 142).383 It is not surprising that Sulis’ temple was attended by priests and
diviners. Unfortunately, these inscriptions provide some of the only evidence of the ritual
which took place at Bath. Consequently, there is little evidence to suggest how many priests
and augurs worked at Bath and no evidence describing what rituals took place at the site
Bath. Consequently much has been left to speculation and comparison with other Romano-
Celtic cult sites.
Bath’s sacred spring was the focal point for votive offerings to Sulis. The spring lay at the
centre of Bath’s temple complex and seemed to be the most acceptable place to approach
Sulis. During the second to fourth centuries CE, the spring received its highest volume of
defixiones.
383 We cannot tell whether or not Lucius worked in Sulis’ temple or was just passing through.
119
The number of defixiones found in Sulis’ spring suggests that Bath’s defigentes must have
believed that the defixiones worked. It has been noted that Sulis continued to be petitioned for
up to two hundred years. To Bath’s lower class, Sulis acted as a supernatural magistrate who
presented a better chance of restitution than an actual magistrate. Sulis’ magisterial role is
emphasized through her persecution of criminals and wrongdoers emphasized through the
curse tablets’ inscriptions.384 The defixiones address Sulis as if they are a petition. A defigens,
most commonly someone who was robbed, asked the goddess to consider the case and to
punish the thief if she thought punishment was deserved. Importantly, Sulis was not
commanded by a mortal. The question of whether or not a thief deserved punishment lay with
Sulis. In this way, Sulis’ judicial qualities are clear.
Conclusion
Although Sulis was clearly the most prominent goddess at Bath, there is limited evidence
describing the site. Consequently, it is difficult to discern what attributes Sulis has a goddess
and which rituals took place at Bath. Due to the size of the temple complex and the elaborate
ornamentation we know that Sulis’ cult and temple complex were prosperous. Sulis’ temple
was built in the Romano-Celtic style emphasizing a degree of hybridization between Roman
and British cultures. Although most of the temple’s iconography and basic layout were
Roman, the temples’ orientation around Sulis sacred spring and the Celtic Gorgon on Bath’s
pediment were an acknowledgment of ancient British culture. Her connection to Bath’s hot
springs has led many to believe that one of Sulis’ functions was the role of a healing goddess.
It is likely that Sulis had a role as a healer at Bath, within Celtic religion the role the deities of
384 For a more elaborate discussion of the defixiones judicial qualities please refer to chapter 1 of this thesis.
120
springs and rivers were primarily female deities.385 As noted above, Celtic water deities have
often been associated with healing attributes. To the Roman mind, Sulis’ connection with a
thermal spring would have given the impression of a potent healing divinity. In addition, the
sheer number of defixiones suggests that they were a central function of Sulis’ cult. The
defigentes at Bath believed they could approach Sulis with their grievances (often theft) and
sought retribution from the goddess. The presence of the defixiones notes that Sulis was
believed to be able to take a life or severely cripple someone. Sulis was able to both heal and
take health away from mortals. In addition, grave stelae attest to the presence of augurs and at
Bath. Bath’s Gorgon pediment as well as inscriptions on dedicated items such as the
defixiones reveals a definite connection with the Roman goddess Minerva. Although Sulis
was conflated with Minerva, she may have meant different things to her Roman and British
worshippers. To the Romans, Sulis may have been a British Minerva Medica with extensive
healing properties. As most of Bath’s defigentes were British, Sulis may have acted as a
judicial divinity they could approach for help. Several scholars such as Green and Cunliffe
have explored the possibility that Sulis was a solar goddess. Although there are
representations of Sol and Luna around Bath, there isn’t enough evidence to draw a solid
connection between the goddess Sulis and Sol. Although we can draw conclusion about
Sulis’ cult and the temple’s ritual, there is still much we don’t understand. Continuing to
study Bath is important for gaining an understanding of Romano-British cults.
385 Green (1996) 90.
121
Conclusion
The Romano-British town of Bath was a prevailing part of Britain’s Roman past. The large
temple complex has provided an example of the aggrandizement of temple complexes in
Britain. The Gorgon pediment has become a powerful symbol of Bath’s Roman history.
Today Bath remains a busy city still centred on the temple of Sulis-Minerva. The temple was
to worship a powerful goddess Sulis-Minerva who represented a fusion of both British and
Roman religious culture. However, the excavation of Bath has proved to be difficult due to
the modern city limiting the amount of evidence which can be used to research the Roman
town of Aquae Sulis. Furthermore, Bath is only cited in one piece of Classical literature and
an Anglo-Saxon poem. Consequently, there are next to no first-hand accounts of what life
was like at Bath or what actions were taking place at the temple of Sulis-Minerva. Sulis is
most commonly recognized for her thermal hot springs as well the large cache of 130
defixiones discovered in the spring. Consequently, Sulis has been thought to be a healer and
an avenging force. In this thesis I have discussed what Bath may have been like and the type
of goddess which made her home at the centre of Bath.
My first chapter explained the chronology of Bath. Bath is considered a Roman ‘small town’
in England. As a ‘small town’, Bath fitted into a wider range of smaller cities with specific
functions built throughout Roman Britain. Bath’s own function was most probably religious
with the town centred on Sulis’ temple complex and sacred spring. Most of Bath’s physical
remains are of a religious type. However, much of the wider city of Bath still remains
unexcavated and still has the potential to yield more evidence of Bath’s city life. It is likely
that Bath began its life as a Roman temple in c.65 CE as a basic Roman temple complex.
Over the years of habitation, Bath became larger and more prominent as reflected in its
122
growing architecture and various dedicated altars. It is likely that Bath reached the height of
its power as a Roman city in the third century CE. But Bath’s prominence was short lived
since, like the rest of Roman Britain, Bath began to decline after the mid fourth century CE.
By the fifth century, Bath’s temple complex was derelict and ceased to function as a Roman
temple. Due to political discord in Rome and changing religious practices in the ancient
world, Bath became more difficult to keep afloat. Bath’s final decades are captured through
sedimentary levels in the court yard as trapped pottery sherds and discarded waste help to
date Bath’s last years. Furthermore, coin hoards help to date when Bath was at its most
prominent as a site. The palaeography of Bath’s Latin can help place a rough date on the
defixiones and when they were being deposited in Sulis’ spring. The defixiones can be
divided into two separate types of Latin which are Old Roman Cursive and New Roman
Cursive. The Old Roman Cursive tablets make up the large majority of Bath’s cache and
these have been dated by Cunliffe to c.175-275 CE. The deposition of the Old Roman
Cursive tablets occurred during the busiest period of Bath’s existence. The large majority of
tablets were being deposited when features such as the dome were placed over Bath’s sacred
spring. The New Roman Cursive tablets were far fewer than their Old Roman counterparts.
This may reflect the decline in Bath’s temple complex as offerings ceased to be made at the
spring. Although Bath went through a surge in popularity it was not long until it declined like
the rest of Roman Britain during the fifth century CE.
Chapter Two was a discussion of the large cache of defixiones which was recovered from
Sulis’ spring. Bath’s curse tablets have been the focus of speculation as they, like other
British defixiones, differ in tone and subject from what has been considered the norm for
Greco-Roman defixiones. The intention of defixiones was to influence the outcome of an
event through invocation of a particular deity or deities. Curse tablets were often used to bind
123
rivals in the courts, in love, and very often in sports. Curse tablets were often very highly
formulaic and developed over time to become intricate in their means of cursing, often
invoking both foreign and local deities to carry out the curse inscribed on the tablet.
However, Bath’s defixiones do have differences to examples associated with the
Mediterranean. The majority of Bath’s defixiones (and those of other British equivalents)
dealt with cases of theft. A defigens would dedicate a curse tablet to Sulis asking the goddess
to recover the stolen item or exact restitution against the thief. All British examples deal with
cases where a defigens has felt wronged by someone and has consequently gone to a deity
where they feel justice would best be received. Due to these differences scholars such as
Versnel have suggested that defixiones such as Bath’s are a form of ‘prayer for justice’ rather
than an outward curse tablet. Versnel’s conclusion does hold some ground. Bath’s defigentes
approach Sulis as supplicants and ask the goddess rather than coerce her into aiding them.
Sulis’ divine status is acknowledged. Furthermore, Bath’s defixiones do include legalistic
language and the thief is given to Sulis as if having been surrendered into her jurisdiction for
punishment. This creates an interesting dynamic between defigens and deity. However, the
line between prayer and cursing is thin in the ancient world and the two may not be
completely mutually exclusive. Often prayers were intended to bring harm against an enemy
similarly to a defixio. Bath’s defixiones shared a lot of consistencies from curse tablets from
the Mediterranean. The gruesome punishments inscribed on the tablets as well as the form of
Bath’s defixiones match their Mediterranean cousins. By rolling up the tablets and inscribing
them backwards, they could not be easily read by a stranger’s prying eyes. The majority of
people dedicating curse tablets at Bath were native Britons. Consequently, although Bath’s
defixiones function similarly to prayers they may also be an ancient British form of a
Mediterranean practice.
124
Chapter Three focused on the goddess Sulis as well as her cult at Bath. Sulis fitted into a
group of British deities who were offered cult on a specifically local basis. Sulis’ power and
identity were tied to the spring and Bath and the city itself. Thus, there is no evidence to
suggest that she was worshipped elsewhere in the ancient world. Furthermore, there is little to
no evidence highlighting explaining Sulis’ cult in pre-Roman Britain. Importantly, most
iconography concerning Sulis is made through a Roman lens. Although Sulis’ temple was
originally built in the form of a classic Roman temple, over time it took on the Romano-
Celtic form which was typical of Roman temples in England. Sulis’ popularity is reflected in
the ongoing process of aggrandizement of the temple and the intricate decorations. On
balance, it is most likely that the ritual and practice at the temple was observed in the Roman
way. Inscriptions provide evidence as to both a priest and augur in ancient Bath. However,
there were also some surviving traditions from the British past captured at Bath. Although
Sulis was joined with the Roman goddess Minerva as ‘Sulis-Minerva’, the name Minerva
always followed that of Sulis. The healing power of Sulis’ spring suited that of Minerva
Medica back at Rome. The healing aspect of Sulis is emphasized by the goddess’ thermal
spring and bathing complex. Furthermore, Bath boasted various Celtic-influenced artistic
features such as the famous gorgon pediment, which boasts various iconographical features
which are associated with the Roman Minerva. Many of Bath’s features such as the
defixiones, on the other hand, do not seem to match the attributes of Minerva as she was not a
goddess often invoked on curse tablets. Furthermore, Sulis received various offerings such as
dedicated altars, coins and paterae. Like most deities, it was likely that Sulis had several
characteristics as a divinity. She may have meant different things to the native Britons and
visiting Romans as she was rooted in both cultures. There is no evidence to suggest that only
a single attribute of Sulis was worshipped at Bath. The problem of minimal evidence means
there is no way to state expressly what kind of practices happened at Bath. Despite Sulis’
125
ambiguity is clear she was an important divinity to the people of Bath and as the city’s patron
deity she held a special place in the heart of Bath.
The patron goddess of the city of Bath is one that is shrouded in obscurity due to a lack of
sources. Several suggestions have been made as to what kind of goddess resided in Bath,
mostly based of Sulis’ connection with the Roman Minerva, the thermal spring and the large
cache of defixiones. It is likely that all of these played an important role in how Sulis’ cult
functioned and all of these say something about the goddess herself. The 130 defixiones
highlight a goddess who was capable of giving harmed devotees restitution for a wrong made
against them. The Gorgon pediment also emphasizes a strong Roman connection with
Minerva. Aquae Sulis, or Bath was clearly an important Roman town in Britain. The large
size of the temple and the rich decoration show its prominence. Britain clearly had a unique
take on the Mediterranean curse tablets and they were adapted to suit the needs of people in
Britain. Although Bath’s tablets could be regarded as prayers they also share a number of
traits in common with defixiones. Britain’s defixiones may be a development on
Mediterranean examples by a native population who thought they had better legal
representation from a deity than a local magistrate. There is still much we do not know about
the temple at Bath and its patron goddess. Thus, the continued study of Bath and Sulis-
Minerva are important when understanding ancient Britain and how it changed during Roman
rule.
126
Bibliography Allason-Jones, L., Billington, S., Green, M. (1996) The Concept of the Goddess. London: Routledge.
Adams, J. N. (1992) ‘British Latin: The Text, Interpretation and Language of the Bath Curse Tablets’, Britannia, 23: 1-26.
Adams, J. N. (2003) Bilingualism and the Latin Language. Cambridge: Cambridge University Press.
Aldhouse-Green M. (2004) ‘Gallo-British Deities and Their Shrines’ in M. Todd, A Companion to Roman Britain. Oxford: Blackwell Publishing Ltd.
Aldhouse-Green, M. (2010) ‘Nobel Savages and Barbarous Enemies’ in Caesar’s Druids Story of an Ancient Priesthood. New Haven: Yale University Press.
Audollent, A. (1904) Defixionum Tabellae. Frankfurt: Minerva GmbH.
Beagrie, N. (1989) ‘The Romano-British Pewter Industry’, Britannia, 20: 169-191.
Blagg, T. F. C. (1979) ‘The Date of the Temple of Sulis Minerva at Bath’, Britannia 10: 101- 107.
Bowman, A.K., (1983) The Roman Writing Tablets From Vindolanda. London: British Museum Press.
Bowman, A. K. (1994) Life and Letters On the Roman Frontier Vindolanda and Its People. London: The British Museum Press.
Bowman, A.K., (1994) Literacy and Power in the Ancient World. Cambridge: Cambridge University Press.
127
Bowman, A.K., Thomas, J.D., (1994) The Vindolanda Writing Tablets. London: British Museum Press.
Brilliant, R. (1974) Roman Art from the Republic to Constantine. London: Phaidon Press Ltd.
Brunn C., Edmondson, J. (2015) The Oxford Handbook of Roman Epigraphy. Oxford: Oxford University Press.
Burnham B.C., Wacher J. (1990) The Small Towns of Roman Britain. Berkeley, Los Angeles: University of California Press.
Burriss, E. E. (1930) ‘The Magic Elements in Roman Prayers’, Classical Philology, 25: 47-55.
Caesar. The Gallic War. (2004) (ed. trans. H. J. Edwards). Cambridge, MA: Harvard University Press.
Carpino, A. A. (2003) Discs of Splendor The Relief Mirrors of the Etruscans. Wisconsin: The University of Wisconsin Press.
Cicero. The Nature of the Gods. (1997) (trans. P.G. Walsh) Oxford: Oxford University Press.
Cleary, A. S. E. (1989) The Ending of Roman Britain. London: B.T. Batsford Ltd.
Cleary, A. S. E. (2004) ‘Britain in the Fourth Century’ in M. Todd. A Companion to Roman Britain. Oxford: Blackwell Publishing Ltd.
Collingwood R. G, Richmond I. (1969) The Archaeology of Roman Britain. Methuen & Co Ltd.
Collingwood, R., Wright, R. P. (1965) The Roman Inscriptions of Britain Volume I Inscriptions on Stone. Oxford: The Clarendon Press.
128
Collins, D. (2008). Magic in the Ancient Greek World. Oxford: Blackwell Publishing Ltd.
Cunliffe, B. (1979) The Celtic World. Maidenhead: McGaw- Hill Book Company Ltd.
Cunliffe, B. (1984) Roman Bath Discovered. Routledge & Kegan Paul.
Cunliffe, B. (1988) The Temple of Sulis Minerva at Bath Volume 2: The finds from the Sacred Spring. Oxford: Oxford University Committee for Archaeology.
Cunliffe, B., Davenport, P. (1985) The Temple of Sulis Minerva at Bath Volume 1: The Site. Oxford: Oxford Committee for Archaeology.
Dark, K. R. (1994) Civitas to Kingdom British Political Continuity 300-800. Leicester: Leicester University Press.
De Grummond, N. T. (1982) A Guide to Etruscan Mirrors. Florida: Archaeological News, Inc.
Dickie, M. W. (2001) Magic and Magicians in the Greco- Roman World. London: Routledge.
Dietrich, B. C. (1993) ‘Curse Tablets and Binding spells: Curse Tablets and Binding Spells from the Ancient World by John G. Gager’, The Classical Review, New Series, 43: 313- 314.
Egelhaaf-Gaiser, U. (2007) ‘Roman Cult Sites: A Pragmatic Approach’, In J. Rupke’s A Companion To Roman Religion. Oxford: Blackwell Publishing. 205-221.
Ekroth, G. (2009) ‘Altars’ in M. Gagarin, The Oxford Encyclopaedia of Ancient Greece and Rome. Oxford: Oxford Printing Press.
Faraone C. A. (1991) ‘The Agonistic Context of Early Greek Binding Spells’ in Faraone, C.A., Obbink D. Magicka Hiera: Ancient Greek Magic and Religion. New York: Oxford University Press.
129
Faraone, C. A. & Obbink, D. (1991) Magika Hiera: Ancient Greek Magic and Religion. New York: Oxford University Press, Inc.
Fear, A. T. (1992) ‘Bladud: The Flying King of Bath’, Folklore, 103: 222-224.
Ferguson, J. (1989) Among The Gods: An Archaeological Exploration of Ancient Greek Religion. London: Routledge.
Flint, V., Richard, G., Luck, G., Ogden, D. (1999) The Athlone History of Witchcraft and Magic in Europe Volume 2: Ancient Greece and Rome. London: The Athlone Press.
Fulford, M. (1973) ‘The Distribution and Dating of New Forest Pottery’, Britannia, 4:160-178.
Gager, J. G. (1992) Curse Tablets and Binding Spells from the Ancient World. New York: Oxford University Press, Inc.
Geoffrey of Monmouth. The History Of the Kings of Britain. (2007) (trans. Neil Wright) Woodbridge: The Boydell Press.
Gerrard, J. (2007) ‘The Temple of Sulis Minerva Bath and the End of Roman Britain’, The Antiquaries Journal, 87: 148-164.
Gerard, J. (2013) The Ruin of Roman Britain an Archaeological Perspective. Cambridge: Cambridge University Press.
Graf, F. (1991) ‘Prayer in Magical and Religious Ritual’ in Faraone C.A., Obbink D. Magika Hiera Ancient Greek Magic and Religion. New York: Oxford University Press. p. 188- 213.
Graf, F. (1997) Magic in the Ancient World. Cambridge: Harvard University Press.
Green, M. (1986) The Gods of the Celts. Gloucester: Alan Sutton Publishing Ltd.
130
Green, M. (1991) The Sun-Gods of Ancient Europe. London: B.T Batsford Ltd.
Green, M. (1996) Celtic Goddesses: Warriors, Virgins and Mothers. New York: George Braziller.
Green, M. (1996) Celtic Art Symbols & Imagery. New York: Sterling Publishing Co., Inc.
Haselgrove, C. (2004) ‘Society and Polity in Late Iron Age Britain’ in M. Todd, A Companion to Roman Britain. Oxford: Blackwell Publishing Ltd. p. 12-29.
Hassall, M. W. C., Wright, R. P. (1973) ‘Roman Britain in 1972’, Britannia, 4: 270-337.
Hatcher, J, Barker, T. C. (1974) A History of British Pewter. London: Longman Group Ltd.
Henig, M. (1994) A Handbook of Roman Art A Survey of the Visual Arts in the Roman World. Oxford: Phaidon Press Ltd.
Henig, M. (1995) The Art of Roman Britain. London: B.T. Batsford Ltd.
Henig, M. (2004) ‘Roman Religion and Roman Culture in Britain’ in M. Todd, A Companion to Roman Britain. Oxford: Blackwell Publishing Ltd. P. 220-241.
Herman J. (2000) Vulgar Latin. Trans. Roger Wright. Pennsylvania: Pennsylvania State Univeristy Press.
Hind, J. (1996) ‘Whose head is on the Bath Temple-Pediment’, Britannia, 27: 358-360.
Inscriptiones Graecae, XII. Inscriptiones insularum maris Aegaei praeter Delum, 1. Inscriptiones Rhodi, Chalces, Carpathi cum Saro, Casi, ed. Friedrich Hiller von Gaertringen. Berlin 1895.
Janowitz, N. (2001) Magic in the Roman World. London: Routledge.
131
Jenkyns, R. (2013) God, Space and City in the Roman Imagination. Oxford: Oxford University press.
Jones, M. E. (1996) The End of Roman Britain. London: Cornell University Press.
Jones, M (2004) ‘Cities and Urban Life’ in M. Todd, A Companion to Roman Britain, Oxford: Blackwell Publishing Ltd.
Jordan, D. R. (1975) ‘A Curse Tablet from a Well in the Athenian Agora’, Zeitschrift fur Papyrologie and Epigraphik, 19: 245-248.
Jordan, D.R. (1985) ‘A survey of Greek defixiones not included in the Special Corpora’, Greek, Roman and Byzantine Studies, 26: 151-197.
Jordan, D. R. (2003) ‘“Remedium Amoris”, a curse from Cumae’, Mnemosyne, 56: 666- 679.
Lattimore, R. (1962) Themes in Greek and Latin Epitaphs. Urbana: University of Illinois Press.
Lewis, M. J. T. (1966) Temples in Roman Britain. Cambridge: Cambridge University Press.
Liversidge, J. (1968) Britain in the Roman Empire. London: Routledge and Kegan Paul Limited.
Livy. Ab urbe condita. (1963) (ed. trans. B. O. Foster) Cambridge, Massachusetts: Harvard University Press.
Lloris, F. B. (2015) ‘Latin Epigraphy: The Main Types of Inscriptions’ in Bruun, C., Edmondson, J. The Oxford Handbook of Roman Epigraphy. Oxford: Oxford University Press.
Luck, G. (2000) Ancient Pathways & Hidden Pursuits: Relgion, Morals and Magic in the Ancient World. Michigan: The University of Michigan Press.
132
Lewis, M. J. T. (1966) Temples in Roman Britain. Cambridge: The University Printing Press.
Luck, G. (2006) Arcana Mundi Magic and the Occult in the Greek and Roman Worlds: A Collection of Ancient Texts. Baltimore: The John Hopkins University Press.
Mallory, J. P., Adams, D. Q. (2006) The Oxford Introduction to Proto Indo European and the Proto Indo European World. New York: Oxford University Press.
Marcus Terenti Varronis. Rerum Rusticarum. (1934) (trans. W. D. Hooper, H. B. Ash) Cambridge, Massachusetts: Harvard University Press.
Marinatos, N. (2000) The Goddess and the Warrior: The Naked Goddess and the Mistress of Animals in Early Greek Religion. London: Routledge.
Mattingly, D. J. (2011) Imperialism, Power and Identity Experiencing the Roman Empire. Princeton: Princeton University Press.
Mattingly, D. J. (2006) An Imperial Possession Britain in the Roman Empire. London: Allen Lane.
Ogden, D. (1999) The Athlone History of Witchcraft and Magic in Europe Volume 2: Ancient Greece and Rome. London: The Athlone Press.
Ogden, D. (2002) Magic, Witchcraft, and Ghosts in the Greek and Roman Worlds. New York: Oxford University Press, Inc.
Ogden, D. (2013) Drakon Dragon Myth and Serpent Cults in the Greek and Roman World. Oxford: Oxford University Press.
Orlin, E. M. (2010) Foreign Cults in Rome: Creating a Roman Empire. New York: Oxford University Press, Inc.
133
Petronius. Satyricon (1996) (trans. P. G. Walsh) Oxford: Clarendon Press.
Pliny the Elder. Natural History (2003) (ed. trans. H. Rackman) Cambridge, Massachusetts: Harvard University Press.
Price, P. (1999) Religions of the Ancient Greeks. Cambridge: Cambridge University Press.
Reece, R. (1980) ‘Town and Country: The End of Roman Britain’, World Archaeology, 12.1: 77-92.
Richmond, I. A., Toynbee, J. M. C. (1955) ‘The Temple of Sulis-Minerva at Bath’, The Journal of Roman Studies, 45: 97-105.
Salway, P. (1981) Roman Britain. Oxford: Oxford University Press.
Salway, P. (1993) The Oxford Illustrated History of Roman Britain. Oxford: Oxford University Press.
Scheid, J. (2003) An Introduction to Roman Religion. Indianapolis: Indiana University Press.
Scheid, J. (2007) ‘Sacrifices for Gods and Ancestors’, in J. Rupke’s A Companion To Roman Religion. Oxford: Blackwell Publishing. P. 263- 272.
Sellwood, L. (1980) ‘Notes on the nine Ancient British coins from Bath’, The British Numismatic Journal, 50: 128- 131.
Shearer, A. (1996) Athene Image and Energy. London: Penguin Books Ltd.
Smith, C. C. (1983) ‘Vulgar Latin in Roman Britain: Epigraphic and other Evidence’ in Aufstieg Und Niedergang Der Romishen Welt. New York: Walter de Gruyter. p. 893- 948.
Smith, D.J. (1983) ‘Mosaics’ in Henig, M. A Handbook of Roman Art A Survey of the Visual Arts in the Roman World. Oxford: Phaidon Press Ltd.
134
Smith, W. (1875) A Dictionary of Greek and Roman Antiquities. London: John Murray.
Solinus. De Mirabilibus Mundi (1895).
Stamper J. W. (2005) The Architecture of Roman Temples the Republic to the Middle Empire. Cambridge: Cambridge University Press.
Stephenson, P. (2009) Constantine Roman Emperor, Christian Victor. New York: The Overlook Press.
Strubbe, J. H. M. (1991) ‘Cursed be he that moves my bones’ in Faraone C.A., Obbink D. Magika Hiera Ancient Greek Magic and Religion. New York: Oxford University Press.
Tacitus, Agricola (ed. 2006) (trans. M. Hutton) Cambridge, Massachusetts: Harvard University Press, Inc.
Tacitus. Germania (1942) (trans. Jackson W.) New York: The Modern Library.
Tatum, W. J. (2011) A Caesar Reader: Selections from Bellum Gallicum and Bellum Civile, and from Caesar’s Letters, Speeches and Poetry. Mundelein: Bolchazy-Carducci Publishers, Inc.
Tibullus, Elegies (2012) (trans. A. M Juster) Oxford: Oxford University Press.
Tilley, B. (1973) ‘Varro the Farmer A Selection from the Res Rusticae’ Foxton: University Tutorial Press Ltd.
Todd, M. (2004) A Companion to Roman Britain. Oxford: Blackwell Publishers.
Turcan, R. (1992) The Cults of the Roman Empire. Oxford: Blackwell Publishers.
135
Turner, E. G. (1963) ‘A Curse Tablet from Nottinghamshire’, The Journal of Roman Studies, 53: 122-124.
Tyers, P. (1976) ‘Roman Pottery from the City of London’, Transactions London Middlesex Archaeological Society, 27: 228-244.
Tyers, P. (1996) Roman Pottery in Britain. London: B.T Batsford Ltd.
Velleius Paterculus. Res Gestae Divi Augusti (1979) (trans. Frederick W. Shipley). Cambridge, Massachusetts: Harvard University Press.
Versnel, H. S. (1991) ‘Beyond Cursing: The Appeal To Justice In Judicial Prayers’, In C. Faraone and D. Obbink, Magika Hiera: Ancient Greek Magic And Religion. New York: Oxford University Press. p. 60-106.
Virgil. Georgics (ed. 2001) (trans. H. Rushton Fairclough). Cambridge, Massachusetts: Harvard University Press.
Wacher, J (1974) The Towns of Roman Britain. London: B.T Batsford Ltd.
Walker, D.R., (1988) ‘Roman Coins from the Sacred Spring At Bath’, in B. Cunliffe, The Temple of Sulis Minerva at Bath, II: Finds From the Sacred Spring. Oxford: Oxford University Committee for Archaeology.
Warrior, V. M. (2006) Roman Religion. Cambridge: Cambridge University Press.
Watson, L. (1991) Arae The Curse Poetry of Antiquity. Leeds: Francis Cairns Publications.
Webster, G. (1980) ‘The British Under Roman Rule A Study in Colonialism’, History Today, 30:19-24.
Webster, G (1986) The British Celts and Their Gods Under Rome. London: B.T. Batsford Ltd.
136
Wilk, S. R. (2000) Medusa Solving the Mystery of the Gorgon. Oxford: Oxford University Press.
Wood, I. (2004) ‘The Final Phase’ in M. Todd, A Companion to Roman Britain. Oxford: Blackwell Publishing Ltd.
Wright, P. R, Richmond, A. I. (1958) ‘Roman Britain in 1957: I. Sites Explored: II. Inscriptions’, The Journal of Roman Studies, 48: 130-155.
Wunsch, R (1897) Defixionum Tabellae Atticae.
Wunsch, R. (1898) Sethianische Verfluchtungstafeln aus Rom. Leipzig.
Van der Meer, L.B. (1995) Interpretatio Etrusca Greek Myths on Etruscan Mirrors. Amsterdam: J.C. Gieben, Publisher.
Velleis Paterculus. The Roman History (1924) (trans. F. W. Shipley) Cambridge, Massachusetts: Harvard University Press.
137
Image List.
Figure 1. Map showing the distribution of Romano-Celtic temples in Britain. Image: Mattingly (2011) 225.
138
Figure 2. Map of selected Romano-British cities and their size groupings. Image: Jones (2004) 177.
139
Figure 3. The Romano-Celtic town of Viroconium Cornoviorum, Britain, 125-400 CE. Image: Jones (2004) 165.
140
Figure 4. The Roman City of Londinium. Image: https://commons.wikimedia.org/wiki/File:Map_Londinium_400_AD-en.svg.
141
Figure 5. Tabulation of Capital letters in Bath defixiones, Tablets 1-24.Image: Cunliffe (1988) 91.
142
Figure 6. Tabulation of Old Roman Cursive letters in Bath’s defixiones, Tablets 30-54. Image: Cunliffe (1988) 92.
143
Figure 7. Tabulation of Old Roman Cursive letters in Bath’s defixiones, Tablets 55-81. Image: Cunliffe (1988) 93.
144
Figure 8. Tabulation of New Roman Cursive letter forms in Bath’s defixiones, Tablets 5, 64-107. Image: Cunliffe (1988) 94.
145
Figure 9. Defixiones found in Caerleon, RIB 323, Wales. C. 43-150 AD. Image: Collingwood, Wright, (1965).
147
Figure 11. Map showing the distribution of Pewter manufacturing scrap wares in Roman Britain. Image: Beagrie (1989) 177.
148
Figure 12. Tabulation of letter forms from the Vindolanda tablets. Image: Bowman, Thomas (1994) 53.
152
Figure 16. Altar Corners of Sulis Minerva portraying Bacchus and an unknown female figure, Temple of Sulis-Minverva, Bath, c. 65-75 CE. Image: Cunliffe (1985) 226.
153
Figure 17. Altar of Sulis portraying Jupiter and Hercules, Temple of Sulis-Minerva, Bath, c.65-75 CE. Image: Cunliffe (1985) 227.
154
Figure 18. Romano- Celtic Gorgon Pediment, Temple of Sulis Minerva, Bath, c. 2nd century CE. Image: Cunliffe (1985) 215.
155
Figure 19. Bath’s Gorgon Pediment with iconography associated with Minerva, Temple of Sulis-Minerva, Bath, c. 2nd century CE. Image: James McBurney.
156
Figure 20. The Façade of the Four Seasons, Temple of Sulis-Minerva, Bath; restoration drawing by S Lysons, Bath. Image: Cunliffe (1985) 213.
157
Figure 21.
Representation of Luna, Temple of Sulis-Minerva, Bath, c. second century CE. Image: James McBurney
158
Figure 22. Bronze head of Sulis Minerva, Temple of Sulis-Minerva, Bath, c. 65-75 CE. Image: http://teakdoor.com/1880136-post11.html.
159
Figure 23. The back of Minerva’s bronze head displaying where the helmet once sat, Temple of Minerva-Sulis, Bath, c.65-75 CE. Image: James McBurney.
160
Figure 24. Romano- Celtic Gorgon Pediment highlighting the ‘Triton’, Temple of Sulis Minerva, Bath, c. second century CE. Image: Cunliffe (1985) 215.
161
Figure 25. The spring of Sulis, Temple of Sulis Minerva, Bath, 2nd century CE onwards. Image: Cunliffe (1985) 37.